SYSTEMATIC IDENTIFICATION

OF MIRNA TARGETS

AND

THE STEPS IN GENE EXPRESSION

REGULATED BY MIRNAS

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF CHEMICAL AND

SYSTEMS BIOLOGY

AND THE COMMITTEE ON GRADUATE STUDIES OF

STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

David Gillis Hendrickson

December 2009

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/hh427ys1294

© 2010 by David Gillis Hendrickson. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

James Ferrell, Jr, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

James Chen

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Andrew Fire

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Tobias Meyer

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

iv

Abstract

In the last decade, RNA interference (RNAi), the process by which small RNAs direct

the post-transcriptional silencing of cognate mRNA transcripts has revolutionized

longstanding paradigms about RNA function. In addition, researchers have harnessed

this pathway for experimentally induced gene silencing in what is arguably one of the

most important technological advances in modern biology. Disruption of the RNAi

pathway results in aberrant development, cancer, and embryonic death suggesting that

RNAi is an integral component of eukaryotic gene expression programming. Although

RNAi is a vast and diverse pathway with marked distinctions between species, the

basic organization has been established largely from work carried out in worms, flies,

humans, and mice. Double stranded RNA (dsRNA) inputs are “diced” by the class III

ribonuclease Dicer into small dsRNA intermediates of ~21-22 nucleotides (nt) in

length which are transferred to the RNA induced silencing complex (RISC) wherein

the guide strand is selected and bound to an Argonaute (Ago) family protein. Target

mRNAs are then recruited to RISC through Watson-Crick base pairing to the guide

strand. Silencing of target transcripts can be directed by Ago mediated cleavage, or

through Ago mediated recruitment of factors that induce translational inhibition and

mRNA degradation. MicroRNAs (miRNAs) are the most common class of

endogenous small silencing RNAs. Many of the molecular details of RISC mediated

gene silencing are poorly understood as current models are based on only a few

miRNA:mRNA target pairs. Here, we present a method for systematic identification of

specific miRNA targets. We demonstrate that immuno-affinity purification (IP) of

Argonaute proteins is a viable method for isolating RISC associated miRNAs and

mRNAs for identification using DNA microarrays. The strong enrichment of mRNAs

with binding sites to the experimentally introduced miR-1 and miR-124 in Ago IPs

from human embryonic kidney 293T cells (HEK293T) validates the utility of this

method. Furthermore, mRNAs classified as targets of miR-1 and miR-124 using this

approach behave like bona fide targets in that they exhibit significant down-regulation

at the mRNA level. To learn about the steps in gene expression regulated by miRNAs,

v

we simultaneously measured miR-124 mediated changes in Ago enrichment, mRNA

abundance, and ribosome occupancy and ribosome density for ~8,000 genes. The

translational parameters were used to estimate apparent changes in translational rate

and were collected using standard polysome profiling in tandem with DNA

microarrays and a novel gradient encoding scheme. We found that for the majority of

the miR-124 targets, changes in mRNA concentration and apparent translation rate are

concordant and that ~75% of the estimated change in protein levels could be

accounted for by changes in mRNA abundance. Our data is most consistent with

models of miRNA inhibition of translation initiation. To rule out miRNA mediated

repressive mechanisms that would not be visible to our translational profiling

(concordant reductions in translational initiation and elongation, co-translational

proteolysis) we tested the protein levels for 13 targets by Western blot and found that

our estimated changes in protein were nearly identical to the actual changes for 12/13

of the proteins measured. In addition, we observed a large dynamic range for miR-124

mediated down-regulation of mRNA abundance and apparent translation rate, and

estimated protein abundance demonstrating the versatility of miRNA mediated

regulation. The concordance between miR-124 specific changes in mRNA level and

translation supports a model wherein these two regulatory outcomes are functionally

linked in a sequential process or regulated by the same cis factors.

We have also sought to learn about the RNAi pathway from a Dicer-centric

perspective. We generated a library of Dicer truncations to test the contribution of

Dicer‟s conserved protein domains to in vitro dicing reactions to learn about

potentially interesting in vivo function as well as for increasing the efficacy of in vitro

dicing as a gene silencing tool. We found that the domain of unknown function 283

(DUF283) may be important for proper spacing in dicing reactions and is part of

Dicer‟s “molecular ruler”. In addition we found that the ATPase/Helicase domains

may inhibit Dicer activity and are dispensable for in vitro dicing, but may play a role

in non-canonical substrate recognition.

vi

Acknowledgements

I would like to thank my parents, Connie and Bill Hendrickson.

vii

Table of Contents

Abstract .......................................................................................................................... iv

Acknowledgements ....................................................................................................... vi

Table of Contents ......................................................................................................... vii

Table of Figures ............................................................................................................ xii

Table of Tables ............................................................................................................ xiv

Chapter 1: Introduction ................................................................................................... 1

Background ................................................................................................................ 1

Overview ............................................................................................................................ 1

The RNAi “Triggers” ......................................................................................................... 3

miRNA/siRNA Biogenesis................................................................................................. 7

Dicer: Structure and Function ............................................................................................ 8

Argonaute Proteins and the RISC complex ...................................................................... 13

miRNA Targeting ............................................................................................................. 15

miRNA Regulation ........................................................................................................... 17

Scope of this Work ................................................................................................... 23

Dicer Domain Function .................................................................................................... 23

Systematic Identification of miRNA Targets ................................................................... 23

Steps in Gene Expression Regulated by miRNAs ............................................................ 24

Chapter 2: Potential Roles for Conserved Dicer Domains in in vitro Dicing .............. 25

Abstract .................................................................................................................... 25

Introduction .............................................................................................................. 26

Results ...................................................................................................................... 30

Generation of a Dicer mutant library ............................................................................... 30

ATPase/Helicase domain ................................................................................................. 30

DUF283 domain ............................................................................................................... 30

viii

PAZ .................................................................................................................................. 32

RNAse Domains ............................................................................................................... 32

Domain Requirements for Efficient Dicing and Properly Sized siRNAs ........................ 34

Discussion ................................................................................................................ 38

Materials and Methods ............................................................................................. 41

Primers Dicer Domains and Point mutations ................................................................... 41

Protein Expression ............................................................................................................ 43

In vitro Dicing Assays ...................................................................................................... 43

Chapter 3: Systematic Identification of mRNAs Recruited to Argonaute 2 by Specific

microRNAs and Corresponding Changes in Transcript Abundance ............................ 44

Abstract .................................................................................................................... 45

Introduction .............................................................................................................. 46

Results ...................................................................................................................... 49

A Method for Isolating and Identifying miRNAs and mRNAs Associated With Ago2 .. 49

Effects of Ago2 Overexpression on mRNA and miRNA Profiles ................................... 51

Systematic Identification of mRNAs Regulated by miR-1 and miR-124 ........................ 54

Seed Matches in the 3‟-UTRs of Putative miR-1 and miR-124 Targets .......................... 56

Relationship Between Overrepresentation in Ago2 Immunopurifications and

Underrepresentation in the Bulk mRNA Pool .................................................................. 58

Relationship Between Size and Number of Seed Matches and Overrepresentation in

Ago2 Immunopurifications .............................................................................................. 61

Analysis of Putative Target mRNAs that Lack 3‟-UTR Seed Matches ........................... 63

Estimation of the Number of mRNAs Regulated by miR-1 and miR-124 ....................... 66

Functions of the High Confidence miR-1 and miR-124 Targets ...................................... 66

Using Ago2 Immunopurification Enrichment and mRNA Expression Changes to Assess

Computational Target Prediction Methods ...................................................................... 68

Discussion ................................................................................................................ 70

A Direct Assay to Identify Targets of Specific miRNAs ................................................. 70

ix

Functional Insights into miRNA Targeting and Regulation ............................................. 71

Insights into miRNA-based Regulation From Recent, Related Publications ................... 72

Materials and Methods ............................................................................................. 76

Plasmids and oligonucleotides ......................................................................................... 76

Cell culture and transfection ............................................................................................. 76

Imunoaffinity purification and RNA isolation ................................................................. 76

Western blots, Sypro Staining, and Nucleic Acid PAGE ................................................. 77

Microarray Production and Pre-hybridization Processing ............................................... 78

Sample Preparation, Hybridization and Washing ............................................................ 79

Scanning and Data Processing ......................................................................................... 80

Microarray Analyses ........................................................................................................ 81

Sequence Data .................................................................................................................. 81

Conservation of Seed Match Sites .................................................................................... 81

Sequence Analyses ........................................................................................................... 81

miRNA Target Predictions ............................................................................................... 82

Gene Ontology and Gene-set Analyses ............................................................................ 82

Acknowledgements .......................................................................................................... 82

Supplementary Figures ............................................................................................. 83

Chapter 4: Concordant Regulation of Translation and mRNA Abundance for

Hundreds of Targets of a Human microRNA ............................................................... 87

Abstract .................................................................................................................... 88

Introduction .............................................................................................................. 89

Results ...................................................................................................................... 92

Systematic Identification of mRNAs Recruited to Argonautes by miR-124 ................... 92

Systematic Measurement of mRNA Translation Profiles ................................................ 95

mRNA Recruitment to Argonautes by miR-124 Leads to Modest Decreases in

Abundance and Translation Rate ................................................................................... 100

x

miR-124 Affects Both the Ribosome Occupancy and Ribosome Density of Hundreds of

Targets ............................................................................................................................ 104

The Effects of miR-124 Transfection on Protein Products of miR-124 Targets ............ 108

Concordant Changes in Abundance and Translation of mRNAs Targeted by miR-124

Suggests That These Two Regulatory Outcomes Are Functionally Linked .................. 110

Changes in Abundance and Translation of miR-124 Ago IP Targets with Seed Matches in

3′-UTRs, Coding Sequences, and 5′-UTRs .................................................................... 114

Efficiency of Recruitment to Argonautes by miR-124 Seed Matches Correlates with

Effects on Both mRNA Abundance and Translation ..................................................... 115

Discussion .............................................................................................................. 116

Materials and Methods ........................................................................................... 121

Plasmids and Oligonucleotides ...................................................................................... 121

Cell Culture and Transfection ........................................................................................ 121

Preparation of Beads for Immunopurifications .............................................................. 121

Immunoaffinity Purifications ......................................................................................... 122

Western Blots ................................................................................................................. 122

Preparation of Cell Extracts for Translation Profiling ................................................... 123

Sucrose Gradient Preparation ......................................................................................... 124

Sucrose Gradient Velocity Sedimentation...................................................................... 124

Gradient Encoding .......................................................................................................... 124

DNA Microarray Production and Prehybridization Processing ..................................... 125

DNA Microarray Sample Preparation, Hybridization, and Washing ............................. 125

Scanning and Data Processing ....................................................................................... 127

Microarray Analyses ...................................................................................................... 129

Sequence Data ................................................................................................................ 130

Acknowledgments .......................................................................................................... 130

Supplementary Figures ........................................................................................... 131

Chapter 5: Concluding Remarks ................................................................................ 145

xi

Dicer Domain Function .......................................................................................... 145

Systematic Identification of miRNA Targets and Steps in Gene Expression

Regulated by miRNAs ............................................................................................ 147

Appendix .................................................................................................................... 153

Text S1. miRNA-effector Complexes Appear to Nonspecifically Bind Streptavidin

Coated Dynal Beads. .............................................................................................. 153

Text S2. Enrichment of Seed Matches to Highly-expressed miRNAs in Ago IPs

from Mock Transfected Cells. ................................................................................ 155

Text S3. Relationship Between Ribosome Occupancy in Mock-Transfected Cells

and the Change in Ribosome Occupancy Following Transfection of miR-124. .... 156

Text S4. Evaluation of the Significance of the Correlation between Changes in

mRNA Abundance and Translation of miR-124 Ago IP targets Following

Transfection with miR-124. ................................................................................... 158

References .................................................................................................................. 160

xii

Table of Figures

Figure 1. The Origins and Biogenesis of Small Guiding RNAs[139] ............................ 9

Figure 1. Generation of Dicer Mutant Library ............................................................. 31

Figure 2. Western Analysis of Insect Cells Expressing Dicer Variants ....................... 33

Figure 3. The PAZ and DUF283 Domains are Required for Efficient Dicing in vitro 35

Figure 4. Generation of Diced siRNAs by DPR and FL Dicer .................................... 37

Figure 1. Ago2 Association with Dicer and miRNAs. ................................................. 50

Figure 2. Overexpression of FLAG-Ago2 Does Not Perturb Overall mRNA

Expression or miRNA Expression. ............................................................................... 53

Figure 3. Comparison of mRNA and miRNA Specifically Associated With Ago2 in

the Absence or Presence of miR-1 or miR-124. ........................................................... 55

Figure 4. Significantly Enriched Motifs in 3′-UTRs Targeted to Ago2 by miR-1 and

miR-124. ....................................................................................................................... 59

Figure 5. Relationship Between Overrepresentation in Ago2 IP and Changes in mRNA

Levels Due to miR-1 and miR-124. ............................................................................. 62

Figure 6. Comparison of Expression Changes of mRNAs Containing Seed Matches in

3′-UTRs and Coding Sequences of miR-1 and miR-124 Ago2 IP Targets. ................. 65

Figure 7. Estimation of the Number of miR-1 and miR-124 Targets. ......................... 67

Figure S1. Disassociation of Dicer from Ago2 IPs in 300 mM KCL .......................... 83

Figure S2. The Length and Number of 3′-UTR Seed Match Sites to miR-1 and miR-

124 Correlates With Enrichment in Ago2 IPs. ............................................................. 84

Figure S3. Using Ago2 IP Enrichment and mRNA Expression Changes to Assess

Computational Target Prediction Methods. ................................................................. 85

Figure 1. miR-124 Recruits Hundreds of Specific mRNAs to Argonautes. ................ 94

Figure 2. Systematic Translation Profiling by Microarray Analysis. ........................... 96

Figure 3. Analysis of Ribosome Occupancy and Ribosome Density in HEK293T

Cells. ............................................................................................................................. 99

Figure 4. miR-124 Negatively Regulates the Abundance and Translation of mRNA

Targets. ....................................................................................................................... 102

xiii

Figure 5. miR-124 Ago IP Targets Decrease in Ribosome Occupancy and Ribosome

Density Due to the Presence of miR-124. .................................................................. 106

Figure 6. The Effect of miR-124 Transfection on Protein Production of miR-124

Targets ........................................................................................................................ 111

Figure 7. Concordant Changes in mRNA Abundance and Translation of miR-124 Ago

IP Targets. ................................................................................................................... 113

Figure S1. mRNA Enrichment Profiles in Ago IPs. ................................................... 131

Figure S2. Streptavidin-coated Dynal Beads Weakly Enrich miR-124 Targets After

miR-124 Transfection. ................................................................................................ 132

Figure S3. Polysome Profiles and Doping Control Fits. ............................................ 133

Figure S4. miR-124 Ago IP Targets Are Likely Destroyed, Rather than Deadenylated

and Stored. .................................................................................................................. 135

Figure S5. Relationship Between the Coding Sequence Length and Changes in

Ribosome Occupancy and Ribosome Density of miR-124 Ago IP Targets Following

Transfection of miR-124. ........................................................................................... 136

Figure S6. Relationship Between Ribosome Occupancy in Mock-transfected Cells and

Change in Ribosome Occupancy Following Transfection of miR-124. .................... 137

Figure S7. Significance of the Correlation Between Changes in mRNA Abundance

and Translation of miR-124 Ago IP Targets. ............................................................. 139

Figure S8. Concordant Changes in mRNA Abundance and Translation of miR-124

Ago IP Targets with 7mer 3′-UTR Seed Matches and miR-124 Ago IP Targets that

Lack a 7mer 3′-UTR Seed Match. .............................................................................. 140

Figure S9. Changes in Abundance and Translation of miR-124 Ago IP Targets With

Seed Matches in 3′-UTRs, Coding Sequences and 5′-UTRs. ..................................... 141

Figure S10. Efficiency of Recruitment to Argonautes by miR-124 Seed Matches

Correlates with Effects on Both mRNA Abundance and Translation. ....................... 143

xiv

Table of Tables

Table 1. Enrichment of seed match sites to miR-1 and miR-124 in Ago2 IP targets

(1% local FDR). ............................................................................................................ 57

Table S1. Summary of miR-124 Targets for Western Blot Analysis. ........................ 144

1

Introduction

Background

Overview

A large focus in modern biology has been the description and discernment of the

mechanisms responsible for the regulation of coherent gene expression programs

required for a cell to respond to external stimulation for the purposes of development

and adaptation. In the past, much emphasis was placed on transcription factors and the

mechanisms by which stimuli are converted into coordinated gene expression patterns.

Likewise, some work has dealt with elucidation of the specific control of mRNA

decay rate as the balance of transcription and mRNA degradation define an mRNAs

half life, and to some extent, the abundance of the cognate proteins. As the correlation

between mRNA and protein abundance is inadequate for predictive purposes for non-

structural genes, many biologists have begun to study post transcription regulatory

mechanisms that control mRNA translation rate and protein stability [1,2,3,4,5,6,7].

Indeed, all evidence is very suggestive that regulation of gene expression is a

multidimensional phenomenon wherein genes are targeted at all stages of production

from DNA to protein through a diverse array of mechanisms. However, other than a

few studies, most research on eukaryotic gene expression was conceptually bound by

the paradigm that the role of RNA was more passive than regulatory.

One such discovery ahead of its time was that of the heterochronic lin-4 gene in

Caenorhabditis elegans[8,9]. The lin-4 gene was identified as a regulatory factor

crucial in the down-regulation of the lin-14 protein; an event required for progression

of normal worm development[8,9]. Surprisingly, lin-4 did not encode a protein, but

rather two extremely short transcripts of only 61 and 21 nucleotides (nt) in length that

shared partial complementarity to sequence elements in the 3‟-untranslated region

(UTR) in the transcript of another gene, lin-14[8,9]. A multitude of similar small

RNAs were subsequently discovered justifying the creation of a new class of non-

2

coding RNAs named microRNAs (miRNAs)[10,11,12]. Mutagenesis of these

sequence elements in lin-14 that were complementary to the lin-4 miRNA resulted in

aberrant larval development suggesting that lin-4 regulates lin-14 expression through

an antisense-sense RNA interaction [8,9].

In plants, the phenomenon of co-suppression had been observed wherein the

introduction of transgenes designed to increase the expression pigmentation factors

resulted paradoxically in reduced expression of the factors[13,14]. Likewise, plants

were also found capable of silencing endogenous genes when exogenous copies of the

same genes were introduced as part the genetic payload of an RNA virus in what was

referred to as virus induced gene silencing (VIGS)[15,16,17,18,19]. Although small

RNAs were not specifically implicated in either observation, an RNA mediated

mechanism had been suggested as responsible for the unexplained silencing[17].

The commonality of homology dependent gene silencing shared between these

observations became apparent upon the seminal Fire et al. discovery that an

experimentally introduced double stranded RNA (dsRNA) trigger produces a robust

silencing of genes complementary to the double stranded trigger[20]. The key

breakthrough in explaining these phenomena came in 1998 when Fire and Mello, in

attempt to attenuate the expression of specific genes in C. elegans using antisense

technology, discovered a silencing response to dsRNA that was much larger in

magnitude when compared to introduction of the single stranded sense or antisense

molecules[20]. This observation has led to the discovery and description of a

mechanism broadly referred to as RNA interference (RNAi) that has presented both an

invaluable tool for biological study and a novel paradigm for thinking about the role of

the non-coding genome and RNA in the regulation of gene expression.

An entire field dedicated to understanding the regulatory functions of RNA has grown

in the decade following the recognition of RNAi as a potent form of post

transcriptional gene silencing. From this research, several salient features have

emerged. First, RNAi is a collective term, referring to several highly conserved small

3

RNA guided pathways capable of regulating mRNA stability, translation efficiency

(co-suppression, VIGS, miRNAs) heterochromatin structure, and

transcription[12,20,21,22,23,24,25,26,27,28,29,30,31,32,33].

In general, RNAi is a process unique to eukaryotes, although some bacterial species

express proteins found in RNAi pathways, and some eukaryotes, most notably

Saccharomyces cerevisiae, have not conserved fully functional RNAi pathways[34].

Historically, RNAi has been separated into two distinct phases, an initiation phase

defined by the biogenesis of the silencing agent, and the effector stage wherein the

silencing agent down-regulates the expression of its cognate complementary

targets[35]. In the initiation step, dsRNA triggers from both endogenous and

exogenous sources are processed into small RNAs between 18-26 nucleotides in

length (depending on the organism) by Dicer, a RNase III family ribonuclease

[36,37,38,39,40,41,42,43,44,45]. Following processing, one strand of the Dicer

product is incorporated into the RNA induced silencing complex (RISC) as a

mechanism for guiding the RISC complex to complementary target

mRNAs[46,47,48,49,50,51]. Argonaute (Ago) proteins are present in all forms of

RISC complexes and physically interact with guiding RNAs and are considered

central to the effector phase[48,49,51,52,53]. Upon association with the RISC

complex, mRNAs can be cleaved directly by RISC, translationally stalled, and/or

indirectly degraded through recruitment of common cellular deadenylases, nucleases,

and decapping enzymes[26,29,30,54,55,56,57,58,59,60,61]. For the purposes of

digesting the wealth of research currently describing all the aspects of RNAi, it is

helpful to review the primary focal points of study on the classification of small

triggers of the RNAi pathway and their biogenesis, the proteins involved in RNAi,

small RNA targeting, and the mechanism of their regulation to highlight unanswered

questions and promising avenues.

The RNAi “Triggers”

An important distinction concerning dsRNA triggers and the mechanism of the

silencing induced by the RISC complex lies with the triggering RNA. The manner in

4

which RNAi regulates gene expression is dependent not only on the sequence of the

input RNA, but rather on the complementarity between the small RNA its target that

determines the functional output i.e. direct cleavage versus translational

inhibition/indirect degradation[58,62,63]. Despite this ambiguity, the origin of small

RNA, along with certain structural characteristics and distinct pathways of biogenesis

are important for classifying the general function and biological roles of the three most

prevalent types of small guiding RNAs known to associate with Ago proteins, and

thus are implicated in RNAi: Small interfering RNAs (siRNAs), miRNAs, and piwi-

interacting RNAs (pi-RNAs)[64].

Small interfering RNAs (siRNAs) are ~21-22nt in length and are defined as

originating from the processive cleaving of long bimolecular dsRNA or long RNA

hairpins with significant dsRNA character[43,44,45]. The large size of siRNA

precursors allows for the generation of a diverse library of siRNAs perfectly

complementary, and thus directed, to the same targets. The Fire et al. observation of

RNAi was in fact triggered by siRNAs from exogenous added dsRNA[20]. RNAi

initiated from siRNAs results in the catalytic cleavage of target mRNAs fomenting a

precipitous drop in target mRNA/protein abundance[43,44,45]. Evidence also suggests

that siRNAs are also capable of generating and maintaining heterochromatin domains

and transcriptional silencing by guiding the RISC complexes to nascent

complementary transcripts to which RISC subsequently recruits histone/DNA methyl

transferases to drive heterochromatin formation in fission yeast[23,65,66,67]. The

mysterious gene silencing observed in co-suppression and in VIGS is mediated by

siRNAs generated from dsRNA arising from the added transgenes and RNA viruses

respectively[68]. Originally, all siRNAs were thought to arise primarily from

exogenous sources to combat intrusion from foreign genetic material belonging to

RNA viruses and any would be invaders that made use of dsRNA[68]. However,

endogenous siRNAs (endo-siRNAs) generated from transposons, centrosomes, and

mobile genetic loci, have been found in both plants in animals and are hypothesized to

protect genomic integrity by initiating destruction of the transcripts responsible for

their generation[69,70,71,72,73,74].

5

Due to the remarkable strength of the response to exogenous dsRNA in worms, RNAi

in the form of experimentally induced gene silencing is a powerful genetic tool[35].

The potency of RNAi in worms also highlights a significant divergence from the

mammalian and insect RNAi pathways. The activity of RNA dependant RNA

polymerases (RdRPs) in worms amplifies the silencing through the generation of

secondary siRNAs created from dsRNA made from the target transcript and using the

original siRNAs as primers[35,75,76,77]. Very recently, flies were found to possess an

RdRP involved in transposon silencing and are generally very amenable to the

addition and uptake of long dsRNA that can be processed by Dicer into a pool of

siRNAs[78]. However, long dsRNA over 29nt in length activates the interferon

response and cannot be used to initiate the RNAi pathway in mammalian systems[79].

RNAi has still served as an experimental boon as techniques utilizing RNAi mediated

gene silencing in mammalian systems have been developed for interrogating complex

biological systems through the addition of DNA constructs encoding short hairpin

RNAs (sh-RNAs) that act as Dicer substrates and give rise to siRNAs aimed at

specific genes of interest. Chemically synthesized siRNAs and in vitro generated diced

pools (discussed below) are also used for the induction of experimental gene silencing

in mammals[80,81,82].

Another type of triggering RNA belongs to the miRNA family first described in

worms[12]. miRNAs are small noncoding RNAs similar in length to siRNAs, but

derived from short hairpin precursor transcripts of endogenous origin whose partial

complementary pairing to target mRNAs potentially regulates expression of more than

60% of genes in many and perhaps all metazoans [9,27,59,83,84,85,86] . The 22nt and

61nt transcripts encoded by the lin-4 loci are the first observed examples of a miRNA

and its precursor[9]. The ubiquity of miRNAs and the widespread effect of their

presence can be seen in the evolutionary conservation of their binding sites in target

mRNAs as well as in the number of highly conserved distinct miRNAs; humans have

at least 400 miRNA genes and there are 140 and 110 well annotated miRNAs in flies

and worms respectively[84,86,87]. Thus, in comparison to canonical protein coding

genes, miRNAs account for ~1-2% of the coding transcriptome[87]. miRNA

6

conservation is quite variable[88]. Examples of ancestral miRNA families stretching

as far back as the emergence of the first RNAi pathways are common as are a

multitude of species specific miRNAs[59,83,88,89]. In terms of the magnitude of the

miRNA mediated gene silencing, recent work directly measuring the effect of miRNA

repression on protein abundance found that miRNAs modestly down-regulate the

expression for hundreds of proteins [90,91]. The range of miRNA repression on total

protein output on average was relatively weak, although a few targets did exhibit >4-

fold changes. Despite these reports of weak changes on protein output, disruption of

miRNA expression and/or the RNAi pathway can result in embryonic lethality, cancer,

and various other disease states suggesting that targets that are highly repressed tend to

be very important, or that the aggregate effect of numerous moderate to weak changes

on hundreds of proteins greatly impacts cellular function. [30,92,93,94,95,96,97,98].

However, these studies concentrated on the removal or addition of single miRNAs.

Most likely, cellular miRNAs target overlapping sets of genes and thus work together

to effect larger total changes than those reported for single miRNA:mRNA

interactions.

miRNAs generally share a much smaller degree of base paring with their targets and

can direct translation silencing, direct target cleavage (if there is enough

miRNA:target base pairing), and indirect degradation of target mRNAs. miRNAs are

distinct from siRNAs in that they inhibit expression of genes unrelated to the loci that

encode them, whereas siRNAs generally target the transcripts from which they arose

[9,26,29,31,33,55,87,96,99,100,101,102,103,104,105,106,107,108,109]. In spite of

extensive study, it is unclear what factors contribute to the regulatory outcomes of

miRNA:mRNA interactions or what steps in gene expression miRNAs

control[27,32,99,110,111,112].

The pi-RNAs are most recently discovered and least understood subset of small RNAs

that associate with the RNAi pathway proteins. piRNAs are distinct from both siRNAs

and miRNAs in their protein binding partners, size (24-30nt), biogenesis, and germ

cell specific expression[113].

7

miRNA/siRNA Biogenesis

Although both miRNAs and siRNAs can feed into similar effector machinery, their

biogenesis is distinct. As mentioned, the initiation phase for siRNAs is succinct. They

are generated from long dsRNA from RNA viruses, active transposons, or

experimentally added transgenes that serve as Dicer substrates[69,70,71,72]. Dicer

processes the RNA from these sources into the canonical siRNAs characterized not

only by their short length of ~21nt, but by a 3‟ 2nt overhang and hydroxyl group in

addition to a 5‟ phosphorylation[43,44,45]. The 3‟-2nt overhang signature is especially

crucial for siRNA/miRNA identification and entry into the RNAi pathway. Together,

these modifications earmark Dicer products for the RNAi effector step as these

structural constraints are required for tight binding to proteins in the RISC complex as

revealed by recent crystallographic studies[51,114,115,116,117,118,119]. miRNA

precursors, however, require two cleavages by separate ribonuclease III enzymes prior

to maturation and RISC incorporation[59]. miRNAs begin life as either intergenic or

intronic sequences transcribed by RNA polymerase II into primary microRNAs (pri-

miRNAs). Still in the nucleus, pri-miRNAs associate with the microprocessor

complex made up by the ribonuclease III Drosha and the RNA binding protein

DGCR8. Here, pri-miRNAs are converted into small hairpins between 60-100nt in

length known as precursor microRNAs (pre-miRNAs) also characterized by a 3‟-2nt

overhang recognized by Dicer[12,37,120,121,122,123,124,125,126,127]. Interestingly,

some miRNAs encoded in intronic regions bypass the Drosha step and are spliced out

of young transcripts in a structural format recognizable by Dicer[128,129,130,131].

Following Drosha cleavage, pre-miRNAs are exported from the nucleus into the

cytoplasm by way of the karyopherin Export-5 when in the presence of a RAN-GTP

cofactor. In the cytoplasm, pre-miRNAs encounter Dicer and are cleaved yet again

into the ~21nt miRNAs[120,121,123,132]. miRNAs are processed by Dicer through

the same catalytic mechanism as siRNAs and have similar properties and chemical

modifications[58]. The Dicer protein associates with RISC complex members for

expeditious handoff of newly converted Dicer products into the heart of the effector

8

complex[49,52,133,134,135,136,137,138]. The origins and biogenesis of small

guiding RNAs are outlined in figure 1[139].

Dicer: Structure and Function

Dicer proteins, like Argonautes, are central to RNAi pathways. Dicer belongs to a

class of RNases specifically evolved to recognize dsRNA. The founding member of

this family is the E. coli RNase III enzyme responsible for processing ribosomal RNA.

Current classification systems categorize RNase III enzymes based on the number

RNase III domains: class one proteins have one domain and class II proteins have

two[27]. Both Dicer and Drosha fall into the second class as both contain two separate

RNase III domains. The standard layout for most Dicer enzymes starts with an N-

terminal ATPase/helicase domain, a domain of unknown function 283 (DUF283), a

PAZ domain, two RNase III domains, and a dsRNA binding domain (dsRBD) domain

(Figure 1)[36,37,39,41,126,140,141].

The exact role of the ATPase/Helicase domain is unknown. A comparison of class II

RNase III enzymes in reveals that the Dicer helicase domain is not conserved in all

Dicers and thus might have a dispensable role in standard dicing, yet may confer

regulatory control over activity and specificity. This region is known to be important

for interactions of Dicer with binding partners which may influence Dicers activity

(below)[142]. Also, the addition of ATP to dicing reactions with drosophila Dicer2

stimulates activity whereas no such ATP preference has been observed in experiments

with human dicer[36,38,135,143]. Likewise, a recent study found that upon mutation

of key residues in the ATPase/helicase domain in human dicer, the enzymatic activity

was greatly reduced for short hairpin miRNA precursors with thermodynamically

unstable stems[144]. Interestingly, the activity of the modified dicer protein was

greater for substrates with highly stable hairpin stems[144]. Thus, the ATPase

domains may vary slightly in Dicers evolved to perform different jobs. DCR-1 and

9

Figure 1. The Origins and Biogenesis of Small Guiding RNAs[139]

10

DCR-2 in drosophila possess slightly different ATPase/helicase domain structure, and

carry out processing of miRNAs and siRNAs respectively[133].

The PAZ domain is an important feature of Dicer proteins allowing for specific

recognition of Dicer substrates and products. The primary function of the PAZ domain

is to help proteins in the RNAi pathway recognize and bind to RNAi specific dsRNA

inputs and can be found in both Dicer and Argonaute proteins. Basically, the PAZ

domain folds to form a hydrophobic pocket to accommodate the 3‟2-nt overhang

common to miRNA precursors, miRNAs, and siRNAs by promoting stacking

interactions between the hydrophobic residues of the PAZ side chains and the nucleic

acid sugar rings[115,117,118,119]. Engineered substrates with longer or shorter 3‟

overhangs bind Dicer with much lower affinity[117]. Crystal structure analysis of the

PAZ domain from Giardia intestinalis revealed that 65 angstroms separate the PAZ

domain from the Dicer catalytic. Satisfyingly, 65 angstroms is also the length of a

stretch of dsRNA ~25-27nt long, the exact size of G. intestinalis Dicer products.

These data support a model in which the PAZ domain recognizes the 3‟ 2nt overhang

common to Dicer substrates and anchors there such that the positioning and length of

space between a PAZ domain and the sites of RNA catalysis serve as the “molecular

ruler” of Dicer proteins[118,119]. This model was found to be correct upon further

study that also found that Dicer proteins can even be “reprogrammed” with different

RNA recognition structures that confer altered specificity to the modified

Dicers[118,119]. Thus, Dicers that do not contain any discernable PAZ domains may

have evolved different methods and structures for recognizing and binding their

cognate substrates[118,119].

The domain of unknown function 283 continues to earn its name. The role of this

elusive domain was nearly exposed by the crystallized G. intestinalis Dicer whose

structure implicated the DUF domain as important for supporting the alpha-helix

connecting the PAZ and RNase III domains based on its similarity to the G.

intestinalis N-terminal platform domain (NTPD)[118,119]. A bioinformatic foray into

11

protein sequence conservation found no significant sequence conservation between the

G. intestinalis NTPD domain and the human DUF283 domain[145]. The authors of

the computational analysis argue that although the two domains have similar folding

topologies βββα and αβββα, the predicted secondary structures are not closely related.

Instead, it is hypothesized that the DUF αβββα fold and the predicted structure are

strikingly similar to topologies and secondary structures indicative of dsRNA binding

domains[145]. The authors postulate that perhaps the DUF domain in conjunction with

the canonical dsRNA binding domain helps to sense the guide strand of Dicer dsRNA

substrates either through direct strand selection, or recruitment of as of yet

unidentified auxiliary RNA binding protein[145]. For the meantime, the DUF domain

remains shrouded in mystery.

Thorough research into the molecular details of the RNase III domains and the dicing

mechanism has yielded a relatively complete description of the dicer mediated RNA

catalysis. As per the function of the PAZ domain, Dicer preferentially cuts dsRNA in

an end specific manner, although RNA catalysis, albeit much slower, is possible when

dsRNA ends are modified to prevent dicing[42,146]. Upon binding a dsRNA target,

Dicer orients its two RNase III domains into a intramolecular psuedodimer such that

Asp1320 and Glu1652 from RNase IIIa, and Asp1709 and Glu1813 from RNase IIIb

form a single catalytic site wherein each domain cuts a single strand of the RNA

substrate 2nt apart so as to generate the 3‟ 2nt overhang required for downstream

binding to the Ago proteins[42,118,146]. Studies analyzing the Dicer cleavage

products of mutated Dicers have revealed that the orientation of the PAZ domain and

the 3‟2nt overhang of the substrate ensure that the RNase IIIa domain cuts the strand

terminating with 3‟-hydroxyl group and that the RNase IIIb domain cuts the strand

bearing the 5‟ monophosphorylation[42,146].

Dicer proteins are increasingly becoming recognized as important regulatory nodes for

cell signaling. As the gate keepers of mature miRNA and siRNA concentration, Dicer

proteins are well poised to regulate the overall input/output of RNA mediated gene

silencing. Extensive binding partners of Dicer have been identified and catalogued in

12

C. elegans and humans, many of which have developmental

roles[147,148,149,150,151]. In humans, Dicer interacts with both the transactivating

response binding protein (TRBP) and protein kinase R activator

(PACT)[138,152,153,154,155,156,157]. These proteins are very similar and may both

be homologues of LOQ, a binding partner of the miRNA specific DICER-1 in flies.

Both TRBP and PACT are RNA binding proteins, and despite their sequence

similarity, have opposing effects on the activity of protein kinase R (PKR) strongly

suggesting significant cross talk between the RNAi and PKR pathways[153,154,156].

A simple pubmed search of the terms “Dicer” and “Cancer” underscore the connection

between Dicer, RNAi pathways, and regulated cell growth.

Dicer is also intimately involved in the effector phase of RNAi and associates with

Argonaute and other annotated RISC binding partners. Effector phase roles of Dicer

are described in further detail below.

Beyond its roles in vivo, Dicer has proven an effective tool in vitro. Myers and

colleagues expressed and purified a recombinant version of human Dicer to generate

“diced pools” of siRNAs targeted at genes of interest[81,139]. Excitingly, recombinant

Dicer was functional and found capable of dicing in vitro transcribed dsRNA into

siRNAs on an experimentally useful scale[81,139]. The use of in vitro dicing has since

become an important tool for experimental gene silencing as Diced pools have several

advantages over chemically synthesized siRNAs. First, diced pools are quite diverse

with one 500bp dsRNA trigger generating ~350 individual siRNAs[158]. This

diversity renders siRNA validation of multiple single siRNAs unnecessary with the

much increased likelihood of successful knockdown on the first try[81,139].

Importantly, diced pools have been shown to reduce interfering off target effects from

the silencing of mRNAs with seed matches to the experimentally introduced siRNA.

The use of diced pools reduces this types of noise by virtue of dilution effects. It is

likely that any given siRNA in the pool also has its own list of off target seed effects;

however the relative concentration of any siRNA from the pool is nominally

13

insignificant. Side by side comparisons of the off target effects of single chemically

synthesized siRNAs and diced pools support these advantages[158].

Beyond the utility of directed gene silencing, in vitro dicing provides a method for

dissecting Dicer structure, function, and mechanism. The ease of doping in labeled

substrates or modified Dicer in in vitro dicing reactions makes this an attractive

method of studying Dicer. Indeed, many of the conclusions supporting the currently

accepted model for the dicing mechanism are built from in vitro dicing assays.

Although a significant body of knowledge exists concerning the mechanism by which

Dicer generates siRNAs and miRNAs, unanswered questions about specific domain

function persist.

Argonaute Proteins and the RISC complex

Central to the effector phase of RNAi are the Argonaute proteins which are

evolutionarily conserved in all interference pathways. RISC has many other

components that can differ between species that are hypothesized to modulate the

magnitude and type of gene silencing activity, RISC localization, and mRNA

targeting, but none are as integral to RNAi function as the Argonaute proteins. As

mentioned, Argonaute proteins belong to two distinct groups: the pi-RNA associated

Piwi proteins and the siRNA/miRNA binding Ago proteins[64,87]. The Ago proteins

have four distinct domains: the N-terminal domain, the PAZ domain, the Mid domain

and the Piwi domain in a N-C primary sequence order[49,51,52,116]. Extensive

structural work has defined the PAZ domain as the primary contributor of binding

specificity to Ago affinity to small RNAs cleaved by Dicer as described

above[51,115,116,117]. A bounty of various solved structures and mutational analysis

provide detailed structural insight in Argonaute function

[51,117,159,160,161,162,163]. Bound RNA spans the rest of the Ago domains

through electrostatic interactions between positively charged Ago side chains and the

negatively charged RNA phosphate backbone. The Mid domain also confers

siRNA/miRNA binding specificity with a small dock for the 5‟ monophosphate[163].

Small RNAs bound to Ago are situated to expose the face of the most 5‟ nucleotides

for engagement in Watson-Crick base pairing with target

14

mRNAs[51,117,159,160,161,162,163]. The PIWI domain has been characterized as a

cryptic RNase H domain and is responsible for cleavage of the target mRNA between

nucleotides 10-11 of the guide strand in interactions with near perfect base

pairing[51,163]. The catalytic residues necessary for this “slicing” activity are

conserved only in Ago family members capable of direct target cleavage (Ago 2 in

humans). In addition, a motif called the PIWI-box in the PIWI domain recruits the

Dicer via an interaction with Dicer‟s RNase III domains[151].

The sequence of RISC assembly differs between worms, flies, and vertebrates. The

majority of research into the RISC formation and loading has been carried out in flies

and humans. The following model is based on work from both flies and human with

an emphasis on the mammalian system. Although the primary role of Dicer is in

generating small RNAs earmarked for entry into the RISC complex, Dicer proteins

serve as a bridge between the initiation and effector phases of RNAi[137]. In humans,

RISC is defined by an Argonaute protein (Ago 1, 2, 3 or 4), associated with Dicer and

the HIV transactivating response binding protein

(TRBP)[49,63,134,138,164,165,166,167,168]. These proteins can associate

independently of a dsRNA trigger, allowing for quick handoff from Dicer to Ago upon

successful Dicer processing[136]. Post Dicer cleavage, dsRNA triggers are passed on

to the Ago protein for guide strand selection. The selection of guide strand is based on

the thermodynamic properties of the small RNA duplex[50,136,169,170,171]. Strands

with 5‟ends at the less stable terminus of the duplex are preferred as guide

strands[169,172]. The passenger strand with the stable 5‟ end is cleaved by slicer

competent Agos and jettisoned from the complex[137,173,174,175,176]. Numerous

examples exist of RISC loaded with both the sense and antisense strands of a small

guiding RNA suggesting that duplexes with ends of comparable stability are able to

contribute either strand to mature RISC. It is unclear how passenger strands are

removed from RISC complexes formed with slicer incompetent Argonaute proteins.

Potentially, the characteristic bulges and gaps present in immature miRNA duplexes

contribute enough instability to the complex as a whole to remove the need for

passenger strand cleavage prior to ejection. Although Dicer is found with Ago in

15

mammalian systems, it is dispensable for correct loading whereas in flies,

DCR2/R2D2 and DCR1/Loquacious are required for RISC maturation from a dsRNA

RNA trigger[136,143,155,169,177,178,179,180,181]. Also in flies, a more complex

system exists for sorting siRNAs and miRNAs into the proper RISC complexes as the

two are processed by two different Dicers and feed into slightly different RISCs

complexes. No evidence so far suggests a similar sorting mechanism in vertebrate

RNAi [182]. In vertebrates, most RISC:mRNA interactions are guided primarily by

miRNAs that share imperfect base pairing with their targets bypassing the need for a

sorting mechanism.

An important RISC cofactor in miRNA repression of both translation and mRNA

stability is the protein GW182. A binding partner of Ago proteins, GW182 associates

with loaded RISC and is sufficient to induce translational inhibition and decay when

tethered to mRNAs sans RISC. It is thought that n-terminal domain of GW182 binds

Ago, and that the c-terminal domain promotes inhibition of translation initiation,

deadenylation, and decay through mechanisms that are not fully understood

[55,56,101,183,184,185,186,187,188,189,190,191,192,193,194,195].

miRNA Targeting

An impressive joint effort combining experimental data with bioinformatics has

successfully outlined the basic features miRNAs recognize in targeting mRNAs. First,

a large portion of miRNA target sites are located in 3‟-UTRs of mRNAs, although

sites in coding sequences and the 5‟-UTRs can also reduce target protein levels of

mRNAs[9,60,83,88,89,196,197,198,199]. Sites in the coding region and 5‟-UTR

however are generally less effective than 3‟-UTR miRNA binding sites, most likely as

a result of translating ribosomes blocking RISC association with target mRNAs

[29,83,88,200,201,202,203]. The most important predictive feature in a candidate

miRNA target is a stretch of six to eight nucleotides complementary to the the 5‟-end

of the miRNA “seed region”. These complementary regions in target mRNAs are the

“seed matches”. To a large extent, it is the interaction between the miRNA seed region

and the mRNA seed match that confer the bulk of affinity and specificity to

miRNA:mRNA target pairs [29,60,89,196]. The importance of the seed match is

16

underscored by their conservation as well as the fact in many instances the seed match

alone is sufficient for repression by the cognate miRNA [60,83,89,199,204].

Beyond the seed region, binding between the 3‟ end of miRNAs and target mRNAs

can, in a small minority of cases, contribute to miRNA targeting and can be

compensatory for seed region mismatches or supplementary to normal seed

interactions [83,84,88,196]. Likewise, a very small degree of conservation can be

found in target mRNAs upstream of the canonical seed match[83,84,88,196]. 3‟

compensatory binding can confer more specificity between miRNAs and their targets

in scenarios in which miRNAs with the same or similar seed are expressed

sequentially. Indeed, in C. elegans, the lin-41 transcript is a target of the miRNA let-7.

However, several miRNAs with the same seed as let-7 are expressed earlier in

development. Thus, the let-7 seed matches in lin-41 are slightly mismatched to the

shared let-7 seed, yet have 3‟ compensatory matches to 3‟ sequence unique to let-7 to

ensure proper timing for down-regulation of lin-41[10,92,98,205,206].

To explain the dual capability of Ago proteins to allow both miRNA mediated seed

interactions and siRNA directed Ago2 cleavage, Bartel in 2009 suggests an attractive

seed nucleation model[87]. First, small guide RNAs bind to the Ago null conformation

in a manner that exposes the seed for the purpose of mRNA targeting while recruiting

the rest of the miRNA inward so as to protect the guide strand from scavenging

cellular RNases. Upon finding a target with a seed match, the mRNA and the guide

strand adopt a half turn helix conformation that does not extend past the seed. When

there is sufficiently strong binding between target and guide in the case of perfect base

pairing, the Ago protein is induced to reorganize and free the 3‟ end of the miRNA to

engage in a helix interaction for the length of the miRNA. However, the extended

helix Ago2 conformation promotes cleavage of the target strand breaking the helix and

turning the guide strand back inward as the broken mRNA is released. The

thermodynamic cost of maintaining the Ago protein in a strained conformation to

accommodate more than half a helix and less than a full 2 turn helix with the target

would also explain the gap in conservation of miRNA binding sites in target mRNAs

17

between the seed match and the less conserved 3‟ binding regions. Extremely recent

crystallographic analysis of Thermus thermophilus Ago with DNA strands of varying

length provide strong biophysical evidence for just such a nucleation model[207].

A seminal study showing that miRNAs induce measurable decreases in the abundance

of some of their cognate mRNA targets provided a powerful experimental handhold

for determining what features in mRNA sequences contribute to miRNA targeting[29].

Basically, Lim et al. found that introduction of either the heart-specific miR-1 or the

brain-specific miR-124 into HeLa cells resulted in significant decreases in the

abundance 96 and 174 mRNAs highly enriched for seed matches to the respective

miRNA. The success of this approach led the authors to expanded the study to

incorporate data from 11 distinct miRNAs[29,88]. This data set was then used to

bioinformatically identify additional features beyond the conservation of canonical

seed matching in the primary sequence of mRNAs that increase the chance that a

mRNA will act as a functional target [88]. These data provided the basis for a model

for the effectiveness of each seed match site in 3‟-UTRs of mRNAs for ~450 miRNAs

that can incorporate conservation, but does not require it, so as to identify potential

specie specific miRNA targets (TargetScan 4.0). It has been argued however that

TargetScan is insufficient for predicting targets that are regulated primarily through

translational inhibition as the data set used to generate the predictions was based

reductions in mRNA abundance. Other prediction algorithms have been formulated

that rely upon both similar and distinct theoretical considerations (e.g. mRNA

secondary structure). However, the accuracy, and thus utility of many of the target

prediction algorithms is limited by a paucity of functional data to test their

performance[89,198,199,208,209,210].

miRNA Regulation

A decade of research has firmly established that miRNAs mediate pos-transcriptional

silencing of their cognate targets. In the mRNA:miRNA interactions driven by

extensive base pairing it is clear that miRNAs, like siRNAs induce direct cleavage by

Argonautes competent for nuclease activity[40,63,109,168,211,212,213]. In the more

18

common scenario (for animals), wherein mRNAs are recruited to RISC through

interactions with miRNA seed regions, the combined knowledge of many avenues of

research has yet to accrete into an accurate description of the mechanisms by which

miRNAs inhibit translation and/or promote target mRNA degradation. Much

conflicting evidence exists arguing that miRNAs either reduce translation, target

instability, or both [9,25,26,27,28,29,30,31,33,86,106]. Often, the discussion

surrounding the gene repressive mechanisms of miRNAs is framed as a debate of

translation versus decay. Actually, a large part of the confusion arises from a lack of

an agreed upon mechanism for either miRNA induced fate. No less than six competing

theories have been put forth as to how miRNAs might reduce translational

efficiency[64]. Likewise, miRNA induced changes in mRNA abundance have been

deemed a result of primary directed degradation as well as attributed to indirect effects

of decreased translation and deadenylation[64]. From this mixed bag of evidence three

explanations can account for reported inconsistencies. One model is that miRNAs act

through a myriad of dissimilar mechanisms that are directed by the recruitment of

different RISC cofactors depending on the miRNA:target pair. This model is

consistent with the idea that there is not a consensus mechanism that could easily

accommodate the varied details of miRNA function. Another explanation is that

miRNAs act through a common initial step after which the proteins associated with

RISC can dictate alternate fates for targeted mRNAs in a sequence and/or cellular

context dependant manner. Lastly, the current confusion surrounding miRNA

regulation could be attributed to experimental artifact.

The sheer number of distinct theories based on solid evidence alone argues that

miRNAs regulate gene expression through numerous different, and sometimes

exclusive, mechanisms. The idea the miRNAs reduce target protein abundance

without altering the half life of an mRNA surfaced early in miRNA research with the

observation that the C. elegans miRNA lin-4 lowers the protein concentration of its

target lin-14 without any effect on the lin-14 transcript abundance[8,9,30]. Similar

work in vertebrate systems bolstered this original finding along with additional

research revealing that targeted mRNAs, miRNAs and RISC components localize to

19

P-bodies suggesting that miRNA targets are actively separated from translating

mRNA pools [56,58,60,108,194,195,214,215,216,217,218].

Another theory of miRNA function is rooted in the well documented deadenylation of

miRNA repressed mRNAs[33,55,57,61,102]. Deadenylation shortens polyA tail

length resulting in decreased initiation[219,220]. However, mRNAS that are removed

or blocked from entering the translational machinery are more likely to become

deadenylated[219,220]. To test whether miRNA mediated deadenylation is a miRNA-

specific effect rather than an indirect result of miRNA induced translation inhibition,

the polyA tail length of mRNAs with internal ribosome entry sites (IRES) that are

immune to miRNA translational repression were measured[61]. Even in these

unrepressed target mRNAs, a marked reduction in polyA tail length was noted[61].

Thus, miRNA mediated deadenylation is a potentially a primary mechanism by which

miRNAs inhibit translational initiation. However, mRNAs lacking an mRNA tail are

also susceptible to miRNA mediated repression[33,108,193].

Other studies have argued for an early cap-dependant inhibition of translation

initiation[99,221]. The first line of evidence implicating miRNAs in cap-dependant

blocked initiation arises from the observation that the Argonatue MID domain binds

the EIF4E complex[221]. Sequestration of EIF4E blocks binding to the 5‟ cap of the

targeted mRNA and prevents subsequent mRNA circularization and recruitment of the

40s ribosomal subunit necessary for translation initiation. Indeed, the addition of

excess EIF4E rescues efficient translation. In Drosophila, 40s recruitment is blocked

by active RISC [222,223]. However, evidence from Drosophila suggests that the

region of Argonaute implicated in EIF4E binding was found to actually bind the RISC

cofactor GW182[193].

The last initiation theory suggests later phases in initiation at the stage of 40s-60s

joining are inhibited by miRNAs[224,225]. It was discovered that Argonaute proteins

associate with both EIF6 and the 60s ribosomal subunit[224,225]. The annotated role

of EIF6 is prevention of premature 40s-60s joining[226,227]. Thus the theory was put

20

forth that initiation is halted by a sequestration of 60s available for 40s joining by Ago

proteins. Furthermore, any available 60s associated with Ago is is blocked from 40s

joining by the presence of the Ago bound EIF6. Two additional lines of evidence

support this model: depletion of EIF6 rescues mRNAs from translational torpor in

worms, and ribosomal toeprinting reveals that translation is stalled just prior to 40-60s

joining as evidenced by the presence of the 40s subunits accumulation at the start

codon of miRNA targeted mRNAs in human cell lysate[224,225].

Outside of initiation, it has also been suggested that post-initiation steps of translation

are targeted by miRNAs [30,201,228,229,230]. In one study, translation driven from

IRES transcripts was reported as repressible by a small guide RNA arguing that

initiation is not a requirement for miRNA translational inhibition. The authors also

observed decreased read through at the stop codon on miRNA repressed mRNAs

indicating that miRNAs slow elongation rate. In addition, miRNA induced ribosome

drop off was implicated when it was noted that ribosomes disassociate from repressed

mRNAs at a faster rate than non-targeted mRNAs post treatment with an initiation

blocking agent[229]. Nascent polypeptide degradation was put forth after the

demonstration that miRNAs and repressed mRNAs associate with actively translating

polysomes[228,230].

A contributing factor to the absence of a consensus model (in addition to the myriad of

potential mechanisms for miRNA induced translational repression) is the reported

phenomenon of miRNA mediated mRNA degradation [26,29,33,61,105,106,188].

That miRNAs reduce the stability of their targets really became established after

publication of the first systematic evidence for widespread miRNA mediated

degradation by exogenous miRNAs in human tissue culture[29]. The intrinsic link

between translation and mRNA stability dictates that some portion of the changes in

mRNA abundance measured in response to the introduction of miRNAs are

necessarily an indirect effect of translational inhibition[219,231,232,233,234].

However, degradation of miRNA targets has been observed in vitro in the absence of

active translation[33,61].

21

Despite the extensive body of work describing multiple diverse mechanisms for

miRNA activity, several lines of evidence based on concordant reductions in inferred

translation and mRNA abundance are consistent with at least a common initial

mechanism. Recent work from Baek et al. and Selbach et al found that changes in

mRNA levels correlate with and account for a majority of the lowered protein

expression observed for miRNA targets after the introduction of a miRNA into human

tissue culture[90,91]. However, concordant regulation of mRNA stability and

translational efficiency may also arise from separate mechanisms regulated by the

same cis factors. The theory that miRNA independent features encoded in targeted

mRNAs dictate the outcome of miRNA:mRNA pairing arises from observations of

both concordant and exclusive reductions in translation efficiency and mRNA

abundance [61,99,101,107,235,236,237].

The use of cell free translation assays, reporter constructs, and/or error introduced by

variations in experimental procedures might also explain some degree of the

divergence in the data used to formulate the disparate theories attempting to describe

the process by which miRNAs inhibit gene expression [99,111]. The majority of work

regarding the translational aspect of miRNA function has relied on reporter assays.

The impetus for the use of such constructs is made clear in consideration of the utility

and power provided by transcripts designed to possess a certain number or type of

miRNA binding sites. However, the uses of this technology can subtlety alter the cell‟s

ability to regulate reporter transcripts in the same manner as endogenous mRNAs.

Artificial mRNAs usually have very short 3‟-UTR sequences and thus may lack

important regulatory information that feeds into normal miRNA function. miRNA

regulation may depend on a delicate balance between targets, miRNAs, the RISC

complex and other cofactors. Weak miRNA mediated repression might be

overshadowed reporter construct mRNAs inundating the cell and disrupting the

regulatory stoicheometry. Not surprisingly, it was reported that experimental

procedure significantly alters the degree and type of miRNA regulation. Factors

influencing mRNA fate determination include but are not limited to, method of

22

transfection, type of 5‟cap attached to an mRNA, and even the promoter sequence

driving reporter expression miRNAs [201,238].

Clearly, research employing systematic methodologies for quantifying miRNA

mediated changes in mRNA abundance, translational efficiency and protein

abundance coupled with careful biochemical analysis will be necessary for resolving

some of these seemingly irreconcilable models. As with many scientific debates

characterized by multiple opposing viewpoints, the truth is most likely somewhere in

between.

23

Scope of this Work

Dicer Domain Function

The method of using recombinant human Dicer for the purposes of experimentally

induced gene silencing was developed in the Ferrell lab. Thus, research into the

structure and function of Dicer and its domains is a logical pursuit based on the

expertise and reagents available. Here we measured the contribution of several of

Dicers various domains to dicing activity through the generation of a library of Dicer

point mutations (key residues identified by previous work) and truncations. This

library was used to express and purify mutant versions of Dicer for in vitro assays

designed to quantify relative efficiencies of activity. In addition, the library can be

used to generate stable human cell lines expressing the Dicer variants to explore the

function of the domains in vivo dicing of biological targets, protein-protein

interactions, and effector stage gene silencing.

Systematic Identification of miRNA Targets

As discussed, algorithms designed to predict miRNA targets are commonly based on

features important for miRNA mediated changes in mRNA abundance, seed match

conservation and predicted secondary structure. However, miRNA mediated

reductions in target transcript concentrations do not fully account for changes in

protein abundance introducing the risk of biasing algorithms based on mRNA changes

towards the prediction of targets most susceptible to this specific regulatory fate and

potentially missing targets with sequence features that primarily limit miRNA

regulation to translational inhibition. Conservation based approaches are limited in the

observation that many conserved sites are not functional and many functional sites are

not conserved. Strategies based on the predicted mRNA secondary structure

surrounding seed matches are predicated on the idea that less structure results in

greater site accessibility for RISC, and thus more miRNA mediated down-regulation.

Predictions based on other predictions with no solid grounding in empirical data are

fraught with error and have not proven especially useful[90,239].

24

To rigorously explore the landscape of miRNA targeting without recourse to

conservation and unbiased with respect to regulatory fate of miRNA targets, we

developed a direct biochemical assay to identify miRNA targets. Using over-expressed

affinity tagged human Argounaute 2 as a handle, we immunopurified the RISC

complex, miRNAs, and their target mRNAs from Human Embryonic Kidney (HEK)

293T cells. The immunopurified (IP) RNA was identified using microarray

technology. This tandem IP/microarray provides a comprehensive method for the

unbiased identification of miRNA targets. To test the method, we transfected both

miR-1 and miR-124 to identify mRNAs specifically recruited to RISC in the presence

of either target. In addition we measured the miR-1 and miR-124 changes in mRNA

abundance to correlate Ago2 enrichment with a functional outcome to validate the

approach.

Steps in Gene Expression Regulated by miRNAs

Efforts to make sense of the various mechanisms by which miRNAs are thought to

mediate gene silencing have, in part, been hampered by our technological inability to

measure multiple parameters that influence actual gene expression simultaneously for

all the potentially hundreds of genes a miRNA may target. Simply knowing to what

extent miRNA targets behave similarly with respect to mRNA abundance,

translational efficiency, and protein abundance would help to determine if there is an

underlying mechanistic commonality between targets of the same miRNA.

Here, we employ our immunoaffinity Ago pull down methodology for systematically

identifying the targets of miR-124 and measuring in parallel, mRNA abundance and

two translational parameters, ribosome occupancy and ribosome density for 8,000

genes and ~650 miR-124 targets. The translational measurements were taken using

polysome profiles, DNA microarrays, and a novel gradient encoding scheme. This

strategy allowed us to directly investigate the behavior of miRNA–mRNA target pairs

with respect to both mRNA fate and translation, on a genomic scale.

25

Potential Roles for Conserved Dicer Domains in in vitro

Dicing

Abstract

RNA interference (RNAi) pathways are cellular mechanisms responsible for

recognizing and responding to various short double stranded RNA inputs. Techniques

utilizing the RNAi pathway have moved to the forefront of reverse genetics, allowing

biologists to effectively silence or inhibit gene translation despite an incomplete

mechanistic understanding. Bridging the gap between RNAi initiation and execution,

the type III ribonuclease Dicer represents an interface between RNAi as an

endogenous mechanism regulating gene expression programs, and RNAi as a

powerful, emergent tool. A two-pronged approach that will conjoin mechanistic data

in vitro with functional data in vivo will at once provide a more complete description

of Dicer‟s roles in the mammalian RNAi pathway, while simultaneously maximizing

the efficacy of Dicer based RNAi techniques.

26

Introduction

Comprehensively, the term RNAi refers to a collection of conserved pathways that

utilize small guiding RNAs to induce gene silencing by down-regulating transcription,

mRNA stability, translation, and transposable element activity in most

eukaryotes[9,12,20,21,22,23,24,25,26,27,28,29,30,31,33,65,66,67,86,106]. RNAi is

commonly subdivided into an initiation step and an effector step[35]. For small

interfering RNA (siRNA) mediated interference, initiation begins with the

introduction and transformation of long double stranded RNA (dsRNA) into 21-25nt

siRNAs with 5‟ phosphates and 3‟ 2nt overhangs[43,44,45]. Micro RNA (miRNA)

driven RNAi starts with the expression of primary miRNA (pri-miRNA) transcripts

bound for conversion via the type III ribonuclease Drosha into ~60-100 nucleotide

miRNA hairpin precursors (pre-miRNA) destined for maturation into 19-25nt

miRNAs, characterized by their 5‟ phosphorylation state, 3‟nt overhangs, and

propensity for bulges, and mismatch base

pairing[12,37,120,121,122,123,124,125,126]. Trigger modifications hinge on the

activity of the type III ribonuclease Dicer that “dices” its substrates into RISC ready

products with the correct 5‟ and 3‟ profiles[43,44,45].

Post Dicer processing, both classes of RNA incorporate into the effector RNA induced

silencing complex (RISC) composed of Argonaute and PIWI family (PPD) proteins

and Dicer[49,52,133,134,135,136,137,138]. In mammals, fully active RISC requires a

single stranded trigger bound to an Argonaute protein[46,48,135]. Mature RISC then

engages in target finding, and subsequent binding to cognate mRNA targets. The

outcome of an mRNA interaction with a complementary RISC bound siRNA/miRNA

depends largely on base pairing[58,62,240,241]. Perfect homology results in

Argonaute mediated target mRNA cleavage whereas mismatch base pairing prevents

translation and and/or an indirect reduction in target mRNA

stability[12,25,26,28,29,30,58,62,240,241]. miRNAs bind imperfectly to regulatory

3‟UTRs from one to a multitude of mRNA targets inhibiting translation and promoting

27

degradation via an unknown RISC

mechanism[9,12,25,26,28,29,30,60,83,89,196,197,198,242]. siRNAs primarily target

the transcripts that gave rise to them with a high degree of complementarity inducing

RISC mediated cleavage. It is unclear if RISC composition is influenced by RNA

cargo[40,46,50,242].

Dicer family proteins are characterized by several common conserved domains. The

most common domains are: a dsRBD, two Rnase domains bearing resemblance to

bacterial Rnase III, a PAZ domain, a domain of unknown function (DUF283) and a

DExH helicase ATPase domain[36,37,39,41,126,140,141]. Fortuitously, some

organisms have split Dicer‟s workload, evolving multiple Dicers with differing

combinations of domains having similar though not completely redundant tasks. This

dissemination of function coupled with recent structural studies suggests very

plausible functions for several Dicer domains[133]. Functional comparison of the two

D. melongaster Dicers, DCR-1 and DCR-2 provides clues into the roles of several

Dicer domains. The helicase containing DCR-2 is required for efficient processing of

long dsRNA into siRNAs, whereas the helicase deficient DCR-1 is not, strongly

suggesting a role for the helicase domain in dicing[133]. Later work revealed that

DCR-1 and DCR-2 process different dsRNA triggers into miRNAs and siRNAs

respectively and are involved in a sorting process responsible for delivering each class

of small RNA into RISC complexes formed around Argonaute 2 for siRNAs and

Argonaute 1 for miRNAs[177,182]. Similarly, structural and mutational analysis of

the Argonaute 2 and Dicer PAZ domain have been adduced to hypothesize that the

PAZ domain is an RNA binding domain responsible for PAZ protein specificity for

RNA earmarked for RNAi[51,117,118,160,161,162,207,212,243,244].

At the outset of this study, the mechanism by which the two human Dicer RNase

domains catalyzed the dicing of Dicer substrates was not fully understood. The solved

crystal structure for Aquifex aeolicus bacterial Rnase III, highly similar to both Dicer

RNase domains, prompted the postulation that Bacterial Rnase III functions as an

28

obligate anti-parallel homo-dimer with two separate cut sites responsible for cleavage

of dsRNA into ~9-11nt products[146,245]. From this model it was extrapolated that

Dicer behaves similarly. Little was known about how spacing for the characteristic

22mer is achieved and was mistakenly believed that the PAZ domain was not required

for this phenomenon[146,245].

Once dsRNA triggers have been correctly processed, the effector stage of RNAi is

initiated through a transfer of small RNA from Dicer to the RISC

complex[134,136,137,138,169,246,247,248]. Experiments carried out on both

drosophila egg and embryo extracts agree that DCR-2 and DCR-1 with their

respective binding partners R2D2 and Loquacious are important for pre-miRNA

maturation and efficient RISC maturation, a process defined by guide strand selection

and passenger strand ejection[143,176,178,179,180,249]. That RNAi silencing

mediated by experimentally introduced processed siRNAs with no need of Dicer

processing is much less efficient in lysates generated from DCR-1 or DCR-2 null

mutants suggests that both drosophila dicers play some redundant role in RISC

maturation downstream of dsRNA dicing or miRNA excising[143,178,179,180,249].

In mammals, the lone Dicer protein serves a purpose more than that of initiation as

well. There is evidence that silencing of exogenous reporters in tissue culture cells

lacking Dicer is severely impaired, even when a chemically synthesized siRNA

capable of feeding directly into the effector step is used[250]. In addition, mammalian

Dicer interacts directly with known PPD proteins[134,137,138,151]. In particular,

Argonaute 2 is known to function as the endonuclease in the RISC complex

responsible for siRNA mediated message cleavage. AGO1, in drosophila, promotes

the stability of mature miRNAs and interacts with DCR-1 during or around the

initiation phase of RNAi [52,54,63,114,151]. Other known RISC factors include VIG

and fragile X, in mammals, and Armitage in Drosophila[251].

Dicer plays a role in exogenous RNAi initiated gene silencing as well. Often

chemically synthesized siRNAs, or constructs encoding Dicer substrates are used to

29

avoid activation of the mammalian interferon response, which is sensitive to dsRNA

over 29nt in length[79]. Myers et al in 2003 discovered that recombinant human Dicer

is useful for dicing dsRNA templates in vitro to create pools of siRNAs directed at one

target[81]. In vitro dicing is particularly attractive, requiring no systematic testing and

validation of multiple siRNA sequences. Furthermore, diced pools are effective at

lower concentrations then the synthetic siRNA, preventing adverse cellular responses

from toxicity and siRNA inundation[81,158]. Interestingly, treatment with proteolysis

appears to increase the efficacy of Dicer suggesting the possibility that one or more of

Dicer‟s domains have an inhibitory role in catalytic activity[42].

Indeed, evidence from multiple organisms has accreted into a framework delineating

the basic tenets and tendencies of RNAi. However, much is unclear and there are

inconsistencies. For instance, a very recent study of in vitro Dicer activity used

mutation of putative catalytic residues, sedimentation data, and Dicer‟s penchant for

dsRNA ends, to put forth a dicer pseudo-dimer monomer model[42,146].

As the primary active initiator of RNAi, as well as a key constituent of the effector

complex, Dicer stands uniquely placed to illuminate many aspects of both the

initiation and effector phases of RNAi. Here we have generated a FLAG tagged

library of Dicer point mutations and truncations that will be employed to shed light on

the molecular details of dicing mechanism in addition to the biological roles of human

Dicer. The library will be used to generate the mutated and truncated Dicers using a

Baculovirus expression system. Dicer activity can then be assayed for using in vitro

dicing reactions. The product and yield, size, and sequence can then be used to gain

insight into the functional requirements of efficient and correctly spaced dicing. In

addition, stable Hela cells expressing members of this library will be used to identify

the structural requirements Dicers interactions with both known and novel binding

partners. Luciferase assays to designed to measure the efficiency of RNAi mediated

gene silencing will be employed in these cell lines in order to determine how and if

Dicer is involved in the effector stage of RNAi based on the requirements of specific

Dicer domains and residues.

30

Results

Generation of a Dicer mutant library

The first step in probing Dicer‟s structure and functions was to generate a library of

point mutation and truncations of interesting Dicer domains based on previous work as

well as conservation implicating several regions as previously unrecognized functional

extensions of the annotated domains (Figure 1).

ATPase/Helicase domain

Prior work on the Arabadopsis thalinia DCL-1 and the D. melongaster DCR-1

implicated the ATPAse/Helicase domains of Dicer. Mutations in several key residues

resulted in aberrant development in and attenuated RNA respectively[39,133].

Multiple sequence alignments with Dicer proteins from different organisms revealed

that several of the mutated residues were conserved, or replaced with similar amino

acids in human Dicer. Overall, two changes were designed to mimic the previously

reported mutations with interesting phenotypes: G31R and C473Y were subcloned

into mammalian/and baculovirus expression vectors. In addition, the helicase domain

was sub-cloned into both vectors by itself to test for potential inhibitory/activation

effects on in vitro dicing and as a well to find potential Dicer binding partners

recruited specifically by the ATPase/Helicase domains in vivo. A Dicer truncation

lacking the ATPase/Helicase domain was generated as well.

DUF283 domain

To test for DUF functionality and binding partners, the DUF domain was cloned alone

into mammalian and baculovirus expression vectors in addition to the creation of a

Dicer truncation lacking both the ATPase/Helicase and DUF283 domains.

31

Figure 1. Generation of Dicer Mutant Library

Left: Illustrations of the Dicer truncations subcloned into expression vectors. All were made except for

the three on the bottom. Right:List of the Dicer point mutation generated and the studies that implicated

the residues as important for function.

32

PAZ

Crystallographic analysis of the PAZ domain implicates this domain as important for

Dicer recognition of its substrates and for conferring the correct spacing of the RNase

domains for effective production of products 21-22 nt in length[51,117,118,119].

Accordingly, Dicer mutants based on the key residues required for proper PAZ

function revealed by crystal structure of the Argonaute PAZ domain crystallized with

a short RNA were generated: F292A, Y309A, L337A and C342A[117]. Likewise, the

PAZ domain was sub-cloned in isolation and in addition to a truncated Dicer lacking

the ATPase/HElicase domain, the DUF domain, and the PAZ domain for the same

purposes as described above.

RNAse Domains

Extensive mutational analysis has determined that there are 4 primary residues

necessary for proper dicing: Asp1320 and Glu1652 from RNase IIIa, and Asp1709 and

Glu1813 from RNaseIIIb[146]. The 24 permutations of single, double, triple and

quadruple mutants were generated to probe the mechanism by which Dicer cleaves

dsRNA. At the outset of this study, it was unclear whether or not Dicer functioned as

pseudo-monomer or as an obligate homodimer[42,146,245]. Studies on bacterial

RNase III suggested that one of the Dicer RNase domains might not be functional.

Additionally, gel filtration analysis has shown that Dicer activity tracks with fractions

of proteins ~twice the size of Dicer. Both pieces of evidence suggested that a single

Dicer molecule may not be sufficient for dicing. The RNase domains also have

important roles in protein-protein interactions as evidenced by the Argonaute PIWI

box and Dicer RNase IIIb interaction[151]. The RNase region of Dicer contains two

well defined RNase domains in addition to a strectch of highly conserved sequence

between them. Each RNase domain was sub-clone in isolation as well as in tandem

with this conserved region of unknown function. These clones will be used to identify

additional Dicer binding partners and the in vivo roles of Dicer in RNAi execution.

Unfortunately, the Dicer truncations lacking one or more of the internal domains, yet

with an intact Helicase domain were never generated. Although baculovirus was

33

Figure 2. Western Analysis of Insect Cells Expressing Dicer Variants

FLAG antibody was used to check for expression of Dicer variants in SF9 cells at increasing

multiplicity of infections (MOI) with recombinant baculovius

34

generated to express all of the aforementioned truncations and domains, not all of the

viruses successfully induced soluble expression of the intended domain. Figure 2

outlines the domains that were successfully expressed and subsequently purified. The

library of point mutation, although created was never used to express and purify

protein, nor were stable cell lines ever created.

Domain Requirements for Efficient Dicing and Properly Sized siRNAs

To identify the minimal domain requirements for efficient Dicing, several truncated

versions of Dicer were expressed in and purified from SF9 insect cells using a

baculovirus expression system: full length Dicer (FL), the DUF, PAZ and Rnase

domains of Dicer (DPR), the PAZ and RNase domains (RP) and the Dicer Rnase

domains with the dsRBD (RNab). Although it is known that the RNase domains are

required, it is unclear what role these other domains play in dicing. To this end,

expressed proteins were purified using a FLAG antibody conjugated to an agarose

resin. Purified proteins were quantified and added at equi-molar concentrations to

dicing reactions containing 500bp of dsRNA (GL3) generated from a fragment of a

firefly luciferase with T7 transcription sites at both ends.

After overnight incubation at 37 degrees Celcius, the reactions were separated by

PAGE on 15% native gels to preserve the double stranded character to distinguish

actual products from single stranded RNA. Interestingly, we observed that the RNab

domains alone were not competent for efficient dicing of dsRNA into 21-22 nt long

siRNAs. A smeared pattern of RNA migration larger than 21-22 was noted indicating

incomplete catalysis of the longer 500bp dsRNA substrate. The RP domains appeared

to process dsRNA, however, the average size of ~15nt is significantly smaller than the

typical Dicer products generated by FL Dicer and the DPR domains (Figure 3A&B).

These data suggest that both the DUF283 and PAZ domains are required for Dicer to

efficiently catalyze the generation of siRNAs of the correct size in vitro. Likewise, the

Helicase domain is dispensable for siRNA production.

35

Figure 3. The PAZ and DUF283 Domains are Required for Efficient Dicing in

vitro

(A) Diced RNA products generated from with Dicer truncations in 16 hour overnight in vitro dicing

reactions with a 500bp dsRNA substrate. Reactions were run out on 4% agarose gels and stained with

ethidium bromide. From left to right: siRNA ~21-22nt, input, DPR Dicer, RNab domains.

(B) Diced RNA products generated from with Dicer truncations in 16 hour overnight in vitro dicing

reactions with a radiolabeled 500bp dsRNA substrate. Reactions were run out on 15% PAGE-TBE gels.

From left to right: RP domains (products <21-22nt), full length Dicer (products ~21-22nt).

36

Reports that Proteinase K increases purified Dicer activity, in tandem with

observations that heterogeneous Dicer purifications are more efficient , compel the

notion that dicing (as defined by the production of ~22nt RNA molecules having 5‟

phosphorylations and 3‟ 2nt overhangs), can be carried out, more efficiently by a

truncated version of Dicer. These observations suggest that perhaps some section or

domain of Dicer is inhibitory to dicing.

To test the hypothesis that FL Dicer may be less active than a truncated version

lacking an inhibitory region, we tested its activity in comparison to the DPR domains,

the one truncation we tested competent for dicing. Equi-molar concentrations of FL

and DPR Dicer were incubated with the GL3 dsRNA over a timecourse from 30

minutes to 24 hours. PAGE analysis revealed that although both proteins were capable

of processing the GL3 dsRNA, the DPR unit produced significantly more siRNA than

the FL Dicer at both 12 and 24hours (Figure 4). At the early time points, FL Dicer

appeared to dice more efficiently, although this observation is potentially an artifact

from the low signal to noise ratio for low intensity bands. Interestingly, plotting the

amount of product generated as a function of time for both the DPR and FL Dicer

reveals that although the DPR reaction is fit well with a simple linear relationship, the

best fit line does not intersect the origin, suggesting potentially that the DPR dicing

has an initial rate that is slower than it rate at later time points (Figure 4). These data

support a model in wherein the ATPase/Helicase domains potentially bind to the

RNase domains and inhibit Dicer activity. It is also possible that the DPR cannot

cleave dsRNA substrates lacking 3‟2nt overhangs as efficiently as full length Dicer.

The potential increase in rate might be indicative of DPR acquiring more substrate

after initial inefficient cleavages.

37

Figure 4. Generation of Diced siRNAs by DPR and FL Dicer

Top: Linear fit of diced products generated in nanograms versus time in hours. The blue line and

equation correspond with the activity of DPR Dicer and the red corresponds with FL Dicer.

(A) Time course of in vitro dicing reaction with FL Dicer. The primary band corresponds with 21-22nt

siRNAs used as loading controls for sizing and quantification of band intensity (not shown).

(B) Time course of in vitro dicing reaction with DPR Dicer. The primary band corresponds with 21-

22nt siRNAs used as loading controls for sizing and quantification of band intensity (not shown).

38

Discussion

Unfortunately, after creating the library of Dicer point mutations and domains, several

studies were published that decreased the urgency of performing many of the planned

experiments including but not limited to, a published analysis of Giardia Dicer, two

independent studies overturning the conjecture that Dicer was required for the RNAi

effector stage from Doi et al, and several works delineating a more complete role

Dicer in human RISC assembly as well identification of novel binding partners and

putative homologs of the DCR1 and DCR2 binding partners R2D2 and

Loquacious[48,136,137,138,181]. In addition the Giardia structure work verified that

Dicer in fact does function as a pseudo-dimer capable of two separate cuts on opposite

RNA strands 2nt apart giving rise to the canonical Dicer 3‟-2nt overhang common to

all Dicer products[118,146].

However, using a small panel of Dicer protein truncations, we found potential roles for

the ATPase/Helicase, DUF283, and PAZ domains in in vitro dicing that goes beyond

published work. The observation that the RNase domains in isolation are unable to

effectively cleave dsRNA is in accord with conclusions made from labs carrying out

similar analysis as well as the crystallographic analysis. The novel observation that the

DUF domain may have a significant role in proper dicing is in agreement with the

hypothesis that the DUF domain may serve as a support platform for the “connector

helix” linking the PAZ and RNase domains[118]. Perhaps in the absence of DUF283,

the PAZ domain correctly recognizes and binds to the ends of substrates but the

RNase domains are unable to orient ~22nt away. That the RP domains generate a

moderately uniform product that is too small suggests that the connector loop in the

absence of the supporting DUF283 is slightly compacted resulting in closer spacing

between the PAZ and RNase domains. Further experiments to test this model should

focus on determining if the addition of separately expressed and purified DUF283

introduced into dicing reactions with the RP domains could restore proper dicing.

Rescue of normal Dicer activity would suggest a crucial for DUF283 as part of human

Dicer‟s molecular ruler. Additionally it might be interesting to test if the effect was not

39

a specific structural interaction between the DUF283 and connecter helix, and rather a

steric affect of DUF283 by assaying the size of Dicer products generated by Dicers

with repeated DUF283 inserted into the primary sequence adjacent to the original

DUF283.

However, as asserted by Dlakic, the DUF283 domain might not share significant

conservation with the Giardia connecter helix and is actually a cryptic dsRNA binding

domain involved potentially in post dicing orientation of Dicer products en route to the

RISC complex or with Dicer recognition of its substrates[145]. To test these theories,

it would be interesting to measure the basic affinity for different combinations of

Dicer‟s domains for dsRNAs with different thermodynamic and structural properties

to determine in what interactions certain domains confer additional affinity or

specificity. If DUF283 in isolation has a measurable affinity for dsRNA and a stable

secondary conformation, NMR experiments designed to determine the structure given

its small size of ~300 amino acids might elucidate both the structure and structural

response to bound dsRNA of DUF283.

The observation that the DPR domains are more efficient at producing siRNAs only

after 12 hours potentially suggests a dual role for the ATPase/Helicase domains as

both an inhibitor of dicing as well as a cofactor for the dicing of non-canonical

substrates. It is easy to postulate an inhibitory role for the ATPase/HElicase domain as

the FL simply generates less siRNAs over time compared to the DPR unit which lacks

the ATPAse/HElicase domain. The proposed role as an initiator for the dicing of non-

standard substrates is rooted in the observation of a higher initial activity for FL Dicer

compared to the DPR, similar in concept to the initial lag seen in Dicer processing of

dsRNA substrates with inaccessible ends[42]. Although, the initial relative burst of

activity we measured for the FL Dicer may be artifactual, it is possible that the effect

is real suggesting that the ATPase/Helicase domain might be important for the

recognition and processing of the initial dsRNA substrate we added which lacks a 3‟-

2nt overhangs common to most human Dicer substrates. One explanation is that the

ATPase/Helicase domain expedites the initial binding or unwinding of substrates

40

lacking recognizable 3‟ overhangs and thus is faster at cleavage than the DPR at the

early time points. The DPR unit however does cleave these long dsRNA, but at a

slower rate. The DPR has an ace in the hole so to speak in the fact that diced products

have the same character as optimal Dicer substrates. This would allow the DPR to

build up a higher concentration of substrates it can readily process at a faster rate over

time, if the ATPase/Helicase domain is a bona fide inhibitor of dicing activity. For this

model to work, it is necessary that the increased proclivity for substrates lacking

overhangs conferred by the ATPase/Helicase domain is greater than the intrinsic

inhibitory effect of this domain on dicing general. Testing this model would require

side-by-side comparisons of FL and DPR Dicer fed a panel of standard and non-

standard potential Dicer substrates to ferret out any subtle benefit the ATPase/Helicase

domain provides for the dicing of non-standard substrates. Importantly, it would be

interesting to test a substrate large enough to discern product from substrate, but not so

large as to produce more substrate after cleavage. A ds RNA ~30nt in length with and

without 3‟2nt overhangs would suffice as the 8-9nt and 21-22nt products cannot be

processed by Dicer. If DPR is less efficient at processing substrates without overhangs

we would expect that FL Dicer would generate much more product than DPR when in

the absence of overhangs and and vice versa in their presence. Additionally, purified

ATPase/Helicase, if folded correctly, would dampen the DPR efficiency if both were

incubated with canonical substrates together.

Fortuitously, it was found that the DPR unit expressed much more efficiently in insect

cells than FL Dicer in addition to enjoying similar if not greater functionality (data not

shown). The DPR unit provides an attractive alternative to FL Dicer for the purposes

of in vitro dicing as an experimental tool for directed gene silencing.

41

Materials and Methods

Primers Dicer Domains and Point mutations

PAZ ATAAGA TAGC GGCCGC ATT GAC TTT AAA TTC

GCATATTACCGTTACTAGACTTAGTACACAC

RNase ATAAGA TAGC GGCCGC AGT CCT GTG ATG GCC

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

Helicase ATAAGA TAGC GGCCGC ATG AAA AGC CCT GCT

CATTAGCAGCTGACTGATGCTAGGCTAGTGAT

RNase+PAZ ATAAGA TAGC GGCCGC CAG AGC CCT TCT ATT GGG

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

Helicase+PAZ ATAAGA TAGC GGCCGC ATG AAA AGC CCT GCT

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

DPR ATAAGA TAGC GGCCGC GGT CCA CGA GTC ACA ATC

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

DUF ATAAGA TAGC GGCCGC GGT CCA CGA GTC ACA ATC

CCG CCG CTC GAGA CTC TTC TTC ATC ATG

R1 ATAAGA TAGC GGCCGC CAG AGC CCT TCT ATT GGG

CCG CCG CTC GAGA GTC TTG ATT TAC TAC

R1+ ATAAGA TAGC GGCCGC CAG AGC CCT TCT ATT GGG

CCG CCG CTC GAGA CCT TTT AAT TAC CGG

R2 ATAAGA TAGC GGCCGC TTT GAA AAG AAA ATC

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

R2+ ATAAGA TAGC GGCCGC TCT TAC GAC TTG CAC

GCTTTAGGCTTATAGACGTACGATGCTAGATCCATT

H1 CATTTTGGGACTAACTGCTTCC TCT TTA AAT GGG AAA TGT GAT

CCA G

GT`AAAACCCTGATTGACGAAGG AGA AAT TTA CCC TTT ACA

CTAGGT C

42

H2 CAAGGAATTGGAAGAAAAGAAACAGAAA CTAGAGAAAATTC

GTTCCTTAACCT TCTTTTCTTTGTCTTTGATCTCTTTTAAG

R137e1 GCTAGTGATGGA TTT AAC CTG GTG CGG CTT GAA ATG CTT

GGC GAC

CGATCACTACCT AAA TTG GAC CAC GCC GAA CTT TAC

GAA CCG CTG

R244d1 GCG GCT TGA AAT GCT TGG CGC CTC CTT TTT AAA GCA

TGC CAT CAC

CGCCGAACTTTAC GAACCG CGG AGG AAAAATTTC GTA

CGG TAG TG

R364e1 GCACTTACCCT GAT GCG CAT GCG GGC CGC CTT TCA TAT

ATG AGA AG

TGTGAATGGGA CTA CGC GTA CGC CCG GCG GAA

AGTATATACTCT TC

R4110e1 CAAAAGCAACACA GAT AAA TGG AAA AAA GAT GAA ATG

ACA AAA GAC

GTTTTCGTTGTGTCTATTTACCTTTTTTCTACTTTACTGTTTT

CTG

R537q2 CAATACTATCACT GATTGTTACGTGCGCTTAGAATTCCTG

GGAGATG

GTTATGATAGTGA CTA ACAATG CAC GCG AAT CTT

AAGGAC CCT CTAC

R644d2 CGCTTAGAATTCCTG GGA GCT GCG ATT TTG GAC TAC CTC

ACA ACC

CGAATCTTAAGGACCCT CGACGC TAAAAC CTG ATG

GAGTGTTGG

R7110e2 CAAAGGCCATGGGGGAT ATT TTT AAG TCG CTT GCT GGT

GCC ATT TAC

GTTTCCGGTTACCCCCTA TAAAAATTCAGC GAA CGA

CCACGGTAA ATG

43

Protein Expression

Protein expression of Dicer Domains was carried out using BAC-to-BAC expression

system as per the manufacturer‟s directions (Invitrogen CAT# A11100)

In vitro Dicing Assays

Dicer purification was performed as previously described [81]. Dicing reactions were

set up and carried out as previously describes [81].

44

Systematic Identification of mRNAs Recruited to Argonaute

2 by Specific microRNAs and Corresponding Changes in

Transcript Abundance

David G. Hendrickson1,$

, Daniel J. Hogan2,3,$

, Daniel Herschlag2,*

, James E. Ferrell1,2

and Patrick O. Brown

2,3,*

1Department of Chemical and Systems Biology, Stanford University School of

Medicine, Stanford, California, 2Department of Biochemistry, Stanford University

School of Medicine, Palo Alto, California, United States of America, 3Howard Hughes

Medical Institute, Stanford University School of Medicine, Palo Alto, California,

United States of America

*To whom correspondence should be addressed. E-mail:

pbrown@pmgm2.stanford.edu or herschla@stanford.edu.

$These authors contributed equally to this work.

This chapter was reprinted from:

PLoS One 2008 May 7;3(5):e2126

PLoS journals publish under the Creative Commons Attribution License (CCAL).

No permission is required from the authors or the publishers.

DGH, DJH and POB conceived and designed experiments. DGH and DJH performed

the experiments, analyzed the data, and helped write the manuscript. POB, JEF, and

DH provided general direction, helped with the data analysis and figure reparation,

and helped write the manuscript.

45

Abstract

microRNAs (miRNAs) are small non-coding RNAs that regulate mRNA stability and

translation through the action of the RNAi-induced silencing complex (RISC). Our

current understanding of miRNA function is inferred largely from studies of the

effects of miRNAs on steady-state mRNA levels and from seed match conservation

and context in putative targets. Here we have taken a more direct approach to these

issues by comprehensively assessing the miRNAs and mRNAs that are physically

associated with Argonaute 2 (Ago2), which is a core RISC component. We transfected

HEK293T cells with epitope-tagged Ago2, immunopurified Ago2 together with any

associated miRNAs and mRNAs, and quantitatively determined the levels of these

RNAs by microarray analyses. We found that Ago2 immunopurified samples

contained a representative repertoire of the cell‟s miRNAs and a select subset of the

cell‟s total mRNAs. Transfection of the miRNAs miR-1 and miR-124 caused

significant changes in the association of scores of mRNAs with Ago2. The mRNAs

whose association with Ago2 increased upon miRNA expression were much more

likely to contain specific miRNA seed matches and to have their overall mRNA levels

decrease in response to the miRNA transfection than expected by chance. Hundreds of

mRNAs were recruited to Ago2 by each miRNA via seed sequences in 3‟-untranslated

regions and coding sequences and a few mRNAs appear to be targeted via seed

sequences in 5‟-untranslated regions. Microarray analysis of Ago2 immunopurified

samples provides a simple, direct method for experimentally identifying the targets of

miRNAs and for elucidating roles of miRNAs in cellular regulation.

46

Introduction

MicroRNAs (miRNAs) are ~22 nucleotide non-coding RNAs that regulate protein

production by pairing to appropriate complementary stretches in

mRNAs[25,59,252,253]. Hundreds of miRNAs are encoded in the human genome,

with an estimated 30% of mRNAs possessing conserved miRNA binding sites,

suggesting that miRNA-based regulation is an integral component of the global gene

expression program[83,86]. The importance and functional range of miRNAs is

evident from their widespread occurrence and the diverse and often strong phenotypes

and disease states associated with mutation or altered expression of

miRNAs[30,92,93,94,95,96,97,98]. miRNAs function through formation of a

ribonucleoprotein complex termed the RNA-induced silencing complex

(RISC)[49,59,137,167]. In humans, RISC is minimally composed of a guide miRNA

bound to an Argonaute protein (Ago 1, 2, 3 or 4), along with Dicer and the HIV

transactivating response binding protein

(TRBP)[49,63,134,138,164,165,166,167,168]. Experiments in mice and human cell

lines show that Ago2 is the central RISC component, capable of cleaving target

mRNA when there is perfect miRNA:mRNA

complementarity[40,63,109,168,211,212,213]. However, most miRNA:mRNA

interactions in metazoans have imperfect complementarity[60,89], and it is likely that

an overwhelming majority of miRNA targets are not cleaved by Ago2. In many cases

it is likely that miRNAs repress translation and/or promote decay of their mRNA

targets[9,26,29,31,33,55,96,99,100,101,102,103,104,105,106,107,108].

A combination of experimental and computational approaches has begun to elucidate

how mRNA targets are specifically recognized by miRNAs. From this large body of

work, several salient features of target recognition have emerged. First, it is likely that

most miRNA target sites are located in 3‟-untranslated regions (UTRs) of

mRNAs[9,60,83,88,89,196,197,198,199]. Sites in coding sequences and, in at least

one instance, 5‟-UTR can also lead to decreased protein levels, although they do so

47

less efficiently than sites in 3‟-UTRs[29,83,88,200,201,202]. Second, a stretch of six

to eight nucleotides near the 5‟-end of the miRNA, the “seed region”, are particularly

important for miRNA function[29,60,89,196]. Their importance is underscored by the

fact that the complementary regions are among the most evolutionarily conserved

regions in mRNA targets and in some instances the seed match alone appears

sufficient to confer recognition[60,83,89,199,204].

The observation that miRNAs cause decreases in the abundance of at least some

mRNA targets provides a powerful strategy for determining what features in mRNA

and miRNA sequences contribute to specificity[26,29,31,33,55,96,101,102,104,105].

Recently, Lim et al. found that transfection of each of two miRNAs, heart-specific

miR-1 and brain/kidney-specific miR-124, into HeLa cells led to decreases in

abundance of at least 96 and 174 mRNAs respectively, many of which were likely to

be direct targets as inferred from the enrichment of seed matches in their 3‟-UTRs

(~90% had 6mer seed matches)[29]. The observation that many of these targets had

conserved seed matches in their 3‟-UTRs and that overexpression of the miRNA

induced a muscle-like or brain-like gene expression program, respectively, suggested

many of the apparent targets were physiological, even though miR-1 and miR-124 are

not normally present in HeLa cells. In addition to the 3‟-UTR sites, the authors found

evidence for some targeting to sites in coding sequences. This miRNA

overexpression/microarray approach was subsequently expanded to 11 miRNAs and

used to identify additional features in mRNAs that contribute to changes in target

mRNA levels[88]. These data provided the basis for a model for the effectiveness of

each seed match site in 3‟-UTRs of mRNAs for ~450 miRNAs (TargetScan 4.0).

Other miRNA target prediction methods are based on limited experimental data and

theoretical considerations (e.g. mRNA secondary structure surrounding predicted

sites), but only limited functional data are available to test their

performance[89,198,199,208,209,210].

One limitation of current approaches is that targets are often inferred from changes in

mRNA abundance; however, miRNA-induced decreases in protein levels can only

48

partially be accounted for by changes in mRNA levels, consistent with the view that

miRNAs affect both translation and mRNA decay

[9,26,29,31,33,55,96,99,100,101,102,103,104,105,106,107,108]. In addition,

identifying targets by altering miRNA expression and measuring changes in mRNA

levels returns no information on which targets might be the most important in carrying

out the actual biological processes (e.g. cellular differentiation) and is limited to the

study of the altered miRNA. Conservation is commonly used as a filter to identify

likely targets, but many functional sites are not conserved and many conserved sites

do not seem to be functional[88,254]. Although useful, the existing methods may be

capturing an incomplete and possibly biased subset of miRNA targets.

A direct experimental method to identify miRNA targets that does not rely on any

specific mechanism of regulation, conservation, or the altered expression of specific

miRNAs is required to fully explore the suite of miRNA targets. Here we describe a

simple method that provides quantitative information about which mRNAs are being

regulated by miRNAs in a cell population. We express affinity-tagged Ago2 in Human

Embryonic Kidney (HEK) 293T cells, immunopurify the resulting tagged Ago2

complexes, and identify the associated mRNAs and miRNAs using DNA microarrays.

This Ago2 immunopurification (IP)/microarray approach allows miRNA targets to be

comprehensively identified in an unbiased fashion, and provides a method for

comprehensively assessing the regulation of mRNAs by RISC. In addition, mRNA

targets of particular miRNAs can be identified by comparing the Ago2 IP/microarray

profiles of cells expressing a particular miRNA to the Ago2 IP/microarray profiles of

untreated cells.

49

Results

A Method for Isolating and Identifying miRNAs and mRNAs Associated With

Ago2

Ago2 is a core component of the RISC complex, associating both with miRNAs and

their mRNA targets[63,168]. Thus, immunopurifying Ago2 under the appropriate

conditions might retain associated miRNAs and mRNAs, allowing miRNA targets to

be identified. At the outset of the project no effective antibodies were available for

immunopurifying Ago2 from mammalian cells. We therefore chose to express a

FLAG-tagged Ago2 protein and identify mRNAs and miRNAs that were associated

with it.

HEK293T cells were transfected with an N-terminal FLAG-tagged Ago2 construct,

allowed to express FLAG-Ago2 for 48 h, and lysed (Figure 1A). Whole cell lysates

from transfected cells were mixed with a FLAG antibody resin, resin was then

washed, and RNA bound to the resin was recovered by phenol-chloroform extraction.

To control for nonspecific association of RNAs with the resin, lysates of mock-

transfected cells were subjected to the same affinity purification.

FLAG IPs of FLAG-Ago2 transfected cells were enriched, relative to mock IPs, in the

RISC components Ago2 and Dicer (Figure 1B and 1C), in small RNAs (Figure 1D),

and in total RNA (5-10 fold; data not shown), consistent with successful purification

of the RISC complex and associated RNAs. This enrichment of RNA and Dicer was

lost when IPs were performed with a two-fold higher KCl concentration, whereas

Ago2 enrichment was not affected (Figure S1).

For microarray analysis, total RNA was isolated from crude lysates, amplified, and

labeled with Cy3, and Ago2-associated RNA was obtained from FLAG IPs, amplified,

and labeled with Cy5. The labeled RNAs were profiled by comparative hybridization

to DNA microarrays printed with the Human Exonic Evidence Based Oligonucleotide

50

Figure 1. Ago2 Association with Dicer and miRNAs.

A) Strategy for the systematic isolation and identification of RNA associated with Ago2. (B)

Immunopurification of FLAG protein purified from mock transfected cells (left) and FLAG-Ago2

transfected cells (right). A Dicer antibody (top) and FLAG antibody (bottom) were immuno-reactive

with bands corresponding to the predicted molecular weight of Dicer (~250 kD) and Ago2 (~90 kD).

(C) SYRO ruby protein stain of FLAG immunopurifications from mock transfected cells (left) and

FLAG-Ago2 transfected cells (right). (D) SYBR gold nucleic acid stain of small RNAs (20–40 bp)

isolated from whole cell lysate (left), FLAG immunopurification from FLAG-Ago2 transfected cells

51

(middle) and FLAG purification from mock transfected cells (right). Brackets outline expected

migration of nucleic acids ~21 base pairs in length.

(HEEBO) probe set containing ~45,000 70mer oligonucleotide probes designed to

detect transcripts for almost all known protein-encoding genes, alternatively spliced

transcripts for hundreds of genes, as well as many annotated non-coding RNAs,

mitochondrial-encoded mRNAs, and viral RNAs[255]. Small RNAs enriched by size

fractionation were labeled with Cy dyes and hybridized to microarrays containing

21mer probes designed to detect ~300 known human miRNAs[256].

Effects of Ago2 Overexpression on mRNA and miRNA Profiles

A major concern was that the miRNAs and mRNAs associated with overexpressed

Ago2 might not be completely representative of those normally associated with

endogenous Ago2. Indeed, although cells transfected with the FLAG-Ago2 construct

were normal in appearance and size, they grew at only ~75% the rate of mock-

transfected cells, indicating some perturbation induced by the overexpression of

FLAG-Ago2.

One method for gauging changes in cell physiology is global gene expression

profiling, which is a sensitive way of assessing changes in cell state since most

physiological responses are associated with changes in mRNA expression[257]. As

shown in Figure 2, the mRNA profiles from the Ago2-transfected and mock-

transfected samples were very similar. Hierarchical clustering showed that the Ago2-

transfected samples did not segregate away from the mock-transfected samples

(Figure 2A), consistent with the hypothesis that Ago2-transfection had little effect on

mRNA levels. Using the significance analysis of microarrays (SAM) algorithm[258],

we found only one transcript with significant differential expression between Ago2-

and mock-transfected samples: a ~25 fold enrichment of a CMV IRES sequence

present in the exogenous FLAG-Ago2 expression transcript. Endogenous Ago2

mRNA levels did not change, as measured by a probe designed to detect the 3‟-UTR

52

of endogenous Ago2. Thus, FLAG-Ago2 overexpression had little detectable effect on

mRNA profiles.

We also assessed whether FLAG-Ago2 overexpression altered the cells‟ miRNA

profile. For the 90 miRNAs deemed to be expressed in HEK293T cells, the levels of

expression in the presence and absence of FLAG-Ago2 transfection were very similar,

with only a few miRNAs registering changes of around two-fold between the two

conditions (Figure 2B). Thus FLAG-Ago2 overexpression had little effect on the

miRNAs present.

The striking similarity in the mRNA and miRNA profiles for the Ago2 and mock

transfected cells suggests that Ago2 overexpression does not substantially alter the

global gene expression program at the mRNA level. These results give us confidence

that conclusions drawn from our IPs are also applicable to unperturbed HEK293T

cells.

Ago2 Immunopurifications Contain a Representative Profile of the Cells’

miRNAs and a Specific Subset of Their Total mRNAs

HEK293T cells were transfected with FLAG-Ago2, and the FLAG-Ago2 associated

miRNAs and mRNAs were isolated, amplified, labeled with Cy dyes, and analyzed by

hybridization to HEEBO and miRNA arrays. The microarray data were then analyzed

by SAM. As shown in Figure 3A, 1215 mRNAs were overrepresented in the FLAG

IPs from FLAG-Ago2-transfected cells relative to mock-transfected cells at a local

false discovery rate (FDR)[259] of 1%. Supervised hierarchical clustering showed that

the profiles of mRNAs significantly enriched in 8 FLAG-Ago2-transfections were

clearly similar to each other and distinct from 14 mock transfection profiles (Figure

3A), demonstrating the reproducibility of the association of this subset of mRNAs

with Ago2. The conservatively-estimated ~1200 overrepresented mRNAs presumably

represent messages being actively regulated by miRNAs under basal conditions. Based

upon gene ontology (GO) terms, the Ago2-associated mRNAs were

53

Figure 2. Overexpression of FLAG-Ago2 Does Not Perturb Overall mRNA

Expression or miRNA Expression.

(A) Unsupervised hierarchical cluster of mRNA levels in HEK293T cells determined relative to a

universal reference for RNA from FLAG-Ago2 transfected cells (red) and mock transfected cells (blue).

Rows correspond to 12,931 gene elements (representing ~9,059 genes) and columns represent

individual experimental samples (rave = 0.92, Pearson correlation between averaged values from each

side of the highest node in the dendrogram). (B) Scatter plot of the normalized log2 microarray signal

intensity of 90 expressed miRNAs from whole cell lysates of mock transfected cells (x-axis) versus the

normalized log2 microarray signal intensity from Ago2 transfected cells (y-axis, r = 0.98). Values are

the averages of 3 experiments. The grey lines delineate the boundary for a two-fold change.

54

diverse in their functional themes, highlighting the likely importance of RISC and

miRNA-mediated regulation in diverse cellular processes.

We also compared the FLAG-Ago2-associated miRNAs to the overall pool of

miRNAs obtained from FLAG-Ago2-transfected HEK293T cells. For the 90 miRNAs

judged as being detectably expressed, their representation in the FLAG-Ago2 IPs

(Figure 3B, y-axis) was proportional to their overall abundance (Figure 3B, x-axis),

with the exception of one miRNA (miR-485-5P) that was about 4-fold down in the

FLAG IP. Thus, FLAG-Ago2 IPs contained a fairly representative array of the cells‟

miRNAs and a select subset of the total mRNAs

Systematic Identification of mRNAs Regulated by miR-1 and miR-124

To identify mRNAs that were recruited to RISC in response to particular miRNAs, we

transfected HEK293T cells with FLAG-Ago2 with or without one of two miRNAs not

normally expressed in HEK293T cells (miR-1 or miR-124), and assessed how the

Ago2-associated mRNA profile was affected . We then used SAM to test for mRNAs

whose association with Ago2 was significantly higher in the Ago2-plus-miR-

transfected cells relative to the Ago2-transfected cells.

The transfection of miR-1 and miR-124 promoted the association of distinct sets of

mRNAs with FLAG-Ago2 and presumably RISC. At a stringent 1% local FDR, SAM

identified 68 mRNAs specifically recruited by miR-1 and 419 mRNAs specifically

recruited by miR-124. Fifty-nine and 388 of these mRNAs, respectively, had RefSeq

IDs with 3‟-UTR sequences available. Hierarchical clustering of mRNAs based on

their association with Ago2 in response to each miRNA revealed a distinct,

reproducible target signature for each miRNA (Figure 3C). There was little overlap

between the sets of mRNAs; only three mRNAs were targeted to Ago2 by both miR-1

and miR-124. These data provide strong evidence that each of these miRNAs recruits

a distinct, reproducible set of mRNAs to FLAG-Ago2-containing RISC complexes.

55

Figure 3. Comparison of mRNA and miRNA Specifically Associated With Ago2

in the Absence or Presence of miR-1 or miR-124.

(A) Supervised hierarchical cluster of putative Ago2 targets that are enriched over mock (local FDR

1%) from FLAG purifications of FLAG-Ago2 transfected cells (red) and mock transfected cells (blue).

Rows correspond to 1,215 gene elements (representing ~1,083 genes) and columns represent individual

experimental samples. There is a high correlation between replicate experiments: rave = 0.80 for Ago2

replicates, 0.73 for mock replicates, and −0.070 for all experiments. (B) Scatter plot of the normalized

log2 microarray signal intensities of 90 miRNAs from whole cell lysate (x-axis) graphed against the

normalized log2 microarray signal intensities of miRNAs associated with Ago2 (y-axis, r = 0.92). Four

replicates were performed for each experiment. The grey lines delineate the boundary for a two-fold

change. (C) Supervised hierarchical clustering of putative miR-1 and miR-124 targets enriched over

Ago2 alone (1% local FDR) from FLAG purifications of FLAG-Ago2 transfected cells alone (red) and

FLAG-Ago2 transfected cells with miR-1 (green) or miR-124 (purple). Rows correspond to 667 gene

features (representing ~544 genes) and columns represent individual experimental samples. rave = 0.80

for Ago2 replicates, 0.77 for Ago2+miR-1 replicates, 0.90 for Ago2+miR-124 replicates, and 0.43 for

all experiments.

56

Seed Matches in the 3’-UTRs of Putative miR-1 and miR-124 Targets

In principle, the mRNAs specifically associated with Ago2 in cells transfected with

miR-1 or miR-124 could have been targeted to Ago2 either directly by the transfected

miRNA or by an indirect mechanism; for example, by other miRNAs whose

abundance or activity is enhanced indirectly by miR-1 or miR-124. If the miRNA

specific Ago2 associated mRNAs consisted predominantly of direct targets, we would

expect that many would contain seed matches to the 5‟ ends of the respective

miRNAs. As an initial approach to this question, we examined what fraction of the

high confidence miR-1 and miR-124 targets possessed seed sequences. By the least

stringent definition of a seed match – a six-nucleotide match complementary to

nucleotides (nt) 2-7 or nt 3-8 in the miRNA – 70% of the miR-1 targets and 75% of

the miR-124 targets possessed seed matches, a highly significant (P < 10-6

and < 10-25

,

hypergeometric density distribution) enrichment over would be expected by chance

(Table 1, designated „6mer‟ seed match). For other more stringent definitions of seed

matches (7mer: complementarity at nt 2-8 or complementarity at nt 2-7 plus an A at

target position 1; 8mer: complementarity at nt 2-8 and an A at target position 1) the

percentage of the miR-1 and miR-124 targets with seed matches was lower but still

highly significant (Table 1). This indicates that the majority of miR-1 and miR-124

targets are likely to be direct targets of the miRNAs.

A second approach to the same question was to ask which 6mers were most highly

overrepresented in the 3‟-UTRs of the high confidence miR-1 and miR-124 targets.

For miR-1, the most highly overrepresented 6mer was UUUUUU. This low

complexity sequence is often speciously enriched in small sample sizes because of its

frequent occurrence in 3‟-UTRs. Thus it is likely not specifically associated with miR-

1. The next three most highly overrepresented 6mers were overlapping sequences with

perfect complementarity to positions 1-8 in miR-1 (Figure 4Ai). We also calculated

the frequency with which a perfect match to each of the 16 6mers in miR-1 was found

in the high confidence miR-1 targets. As shown in Figure 4Aii, the three most 5‟

6mers were highly overrepresented. Similarly, the multiple expectation maximization

57

Table 1. Enrichment of seed match sites to miR-1 and miR-124 in Ago2 IP targets

(1% local FDR).

58

for motif elicitation (MEME) motif discovery algorithm identified a 10 nt motif

sequence from the high confidence miR-1 targets with perfect complementarity to

positions 1-8 in miR-1 (Figure 4Aiii)[260].

Analogous results were found for miR-124, although in this case there was some

enrichment for base pairing near the 3‟ end of the miRNA as well as base pairing near

the 5‟ end (Figure 4Bi and Bii). These results are consistent with previous reports

demonstrating no preference for base-pairing immediately adjacent to the seed match,

but some base-pairing with miRNA positions 13-18[88,198]. The 10 nt motif returned

by MEME included a 7 nt stretch with perfect complementarity to positions 2-8 in

miR-124.

We also looked for overrepresentation of the nt 2-7 seed matches for all 90 of the

miRNAs we deemed as being expressed in HEK293T cells in the putative miR-1 and

miR-124 targets. None were found to significantly overrepresented; the response

appeared to be specific.

Taken together, these data indicate that most of the high confidence targets of miR-1

and miR-124 are likely to be direct targets. In addition, the same sequence criteria

inferred from previous studies for the recognition of mRNA targets by miRNAs

[60,88,89,196,197], especially the importance of sequences complementary to

positions 1-8 in the miRNA, emerged independently from analysis of the mRNAs

overrepresented in FLAG-Ago2 IPs. These results provide an important validation of

the IP method.

Relationship Between Overrepresentation in Ago2 Immunopurifications and

Underrepresentation in the Bulk mRNA Pool

miRNAs appear to regulate gene expression by effects on mRNA abundance or

translation or both. Therefore, previous studies focusing on changes in mRNA levels

59

Figure 4. Significantly Enriched Motifs in 3′-UTRs Targeted to Ago2 by miR-1

and miR-124.

(A) Analysis of mRNAs associated with Ago2 from cells transfected with FLAG-Ago2 and miR-1

relative to cells transfected with Ago2 alone (1% local FDR). (i) Enrichment of hexamers in 3′-UTRs of

miR-1 IP targets compared to 3′-UTRs of all mRNAs passing array filters. Shown are hexamers with at

least four contiguous Watson-Crick base pairs to miRNA with a p-value cut-off of 0.001 (binomial test

with bonferroni correction). Rank by p-value relative to all 4096 hexamers. Bases in red can form

60

Watson-Crick base pairs with miR-1. (ii) Moving plot of observed/expected ratios of hexamers

complementary to miR-1. Frequencies calculated as in (i). (iii) 10mer motif returned by MEME motif

finder using 3′-UTR sequences from the miR-1 high confidence target set. For each position in the

motif, the combined height of the bases represents the information content at that position, whereas the

relative heights of the individual bases represent the frequency of that base at that position. Bases in red

can form Watson-Crick base pairs with miR-1. Numbers underneath the logo correspond with miRNA

5′-position, with 1 being the 5′-most miRNA nucleotide. (B) Same as in (A), except for mRNAs

associated with Ago2 from cells transfected with FLAG-Ago2 and miR-124 relative to cells transfected

with Ago2 alone.

alone potentially miss many targets. We hypothesized that the IP method could

capture all direct mRNA targets regardless of functional outcome. To test this, we

measured mRNA levels from cells transfected with the respective miRNAs in parallel

to the IP experiments. We found 0 and 145 mRNAs significantly decreased in

presence of miR-1 and miR-124, respectively, at a 1% local FDR threshold. Thus

significantly more putative miRNA targets, 56 and 388 respectively, were identified

by miRNA-specific enrichment Ago2 IPs than from the miRNA-specific changes in

mRNA levels.

To further explore the relationship between Ago2 IP enrichment and mRNA

expression change, we relaxed the stringency for the SAM analysis of miRNA-

induced decreases in mRNA abundance to a 10% local FDR threshold (16 for miR-1

and 255 for miR-124), and mapped the values from each assay onto two axes (Figure

5). We broke the data into three color coded classes: mRNAs that are overrepresented

in Ago2 IPs and decrease in mRNA level (Figure 5, black lines); mRNAs that are

overrepresented in Ago2 IPs but do not significantly decrease in mRNA level (Figure

5, red lines); and mRNAs that are not overrepresented in Ago2 IPs, but decrease

significantly at the mRNA level (Figure 5, blue lines). To enrich for the highest

confidence targets, we focused on mRNAs with a 7mer seed match in their 3‟-UTRs.

Only a minority of the targets that were significantly overrepresented in the Ago2 IPs

were also significantly decreased in their mRNA levels: 24% of the miR-1 IP targets

and 40% of miR-124 IP targets (Figure 5, black lines). These targets are represented

61

by lines that run from the positive side of the Ago2 IP axis to the negative side of the

expression axis (Figure 5, black lines). On average the mRNA levels of these miR-1

and miR-124 targets decreased 58% and 61%, respectively. The majority of the targets

overrepresented in Ago2 IPs did not show significant decreases (10% local FDR) in

their overall mRNA levels (Figure 5, red lines). Nevertheless, approximately 90% of

these IP targets did show some decrease in their expression levels (Figure 5, red lines),

with average decreases of 24% and 25%, respectively. Thus, most of the mRNAs

overrepresented in Ago2 IPs did exhibit a modest decrease in their mRNA levels. The

mRNA targets whose levels decreased only modestly may be regulated primarily at

the translation level or miRNAs may play only small modulatory roles in the

expression of the proteins these mRNAs encode.

Conversely, 44% and 56% of mRNAs that decreased significantly in response to miR-

1 and miR-124 transfection, respectively, were not significantly overrepresented (1%

local FDR) in the Ago2 IPs (Figure 5, blue lines). However, almost all of these (9/9

miR-1 targets and 63/65 miR-124 targets) were enriched to some extent in the Ago2

IPs to some extent (Figure 5, blue lines). This trend argues that most of these mRNAs

are actually miR-1 and miR-124 targets.

Relationship Between Size and Number of Seed Matches and Overrepresentation

in Ago2 Immunopurifications

Bartel and co-workers[88] previously reported that mRNAs with long seed match sites

(e.g. 8mer matches) were more likely to change in abundance in response miRNA

transfection than those with shorter seed match sites, and that mRNAs with two 7mer

seed match sites were more likely to show changes than those with one. We therefore

asked whether the same relationships would hold based on overrepresentation of

mRNAs in Ago2 IPs. As shown in Figure S2, this was indeed the case. For both miR-

1 and miR-124, mRNAs with a single 8mer seed match site were overrepresented in

the Ago2 IPs relative to those with single 7mer seed match sites, and those with single

7mer seed match sites were overrepresented relative those with single 6mer seed

62

match sites (Figure S2A and B). Likewise, mRNAs with two or more 7mer seed match

sites were overrepresented in the Ago2 IPs relative to those with one (Figure S2C and

Figure 5. Relationship Between Overrepresentation in Ago2 IP and Changes in

mRNA Levels Due to miR-1 and miR-124.

(A) Lines connect the log2 of the average Ago2 IP value (bottom axis) to the log2 of the average mRNA

expression change (top axis) for three groups of mRNAs from miR-1 experiments. This analysis has

only mRNAs with 7mer seed matches in their 3′-UTRs. Black lines correspond to mRNAs that were

Ago2 IP targets and decreased at the mRNA level (10% local FDR); seven mRNAs are in this group.

Red lines correspond to IP targets but did not decrease significantly at the mRNA level; 20 of 22

mRNAs in this group decrease (log2 change<0) at the mRNA level (P<10−5

, one-way binomial test).

Blue lines correspond to mRNAs that decreased at the mRNA level, but were not Ago2 IP targets; all 9

mRNAs in this group are overrepresented (log2 enrichment>0) in Ago2 IPs (P = 0.0006). (B) as in (A)

63

except for mRNAs from miR-124 experiments. 82 mRNAs are in black. 109/121 mRNAs in red

decrease at the mRNA level (P<10−15

); and 63/65 mRNAs in blue are up in Ago2 IPs (P<10−15

).

D). The same relationships were found in the changes of mRNA levels in response to

miR-1 and miR-124 (data not shown). These findings corroborate and extend those of

Grimson et al.[88] and further validate the Ago2 IP method.

Analysis of Putative Target mRNAs that Lack 3’-UTR Seed Matches

Interestingly, a large minority, 30% and 25%, of high confidence miR-1 and miR-124

targets identified by Ago2 IP do not contain a 6mer seed match in their 3‟-UTRs.

Several studies have provided evidence that seed matches in the coding sequence and

5‟-UTR can also confer regulation by a miRNA, as judged by mRNA expression data,

Ago2 IPs in Drosophila, phylogenetic conservation analyses and reporter

studies[29,83,88,200,201,202]. We therefore checked for enrichment of 7mer seed

matches in the coding sequences of miR-1 and miR-124 Ago2 IP targets. As reported

in Table 1, 29% and 33% of Ago2 IP targets contained coding sequence 7mer seed

match sites for miR-1 and miR-124 respectively (P = 0.007 and 10-20

, hypergeometric

distribution). Further, 59% and 47% of miR-1 and miR-124 targets that lacked any 3‟-

UTR seed matches contained 7mer seed matches in their coding sequences (P < 10-5

and 10-15

; Table 1). 5‟-UTR 7mer seed matches were also significantly

overrepresented (P = 0.001 and 0.0005 for miR-1 and miR-124, respectively; Table 1).

We next set out to assess the effectiveness of 3‟-UTR seed matches versus coding

sequence seed matches in the Ago2 IP targets, as assessed by effects on mRNA

expression (Figure 6A and B). Seed matches in the 5‟-UTR were not included in this

analysis because of their small numbers. We considered two subsets the miR-1 and

miR-124 targets: those mRNAs that possessed a 7mer seed match in the 3‟-UTR but

no 6mer seed match in their coding sequence (Figure 6, red curves) and those that

possessed a 7mer seed match in their coding sequence but no 6mer seed match in their

3‟-UTR (Figure 6, green curves). As a comparison group we examined all mRNAs

(not just those whose levels in the Ago2 IPs changed after miRNA transfection) that

64

contained no 6mer seed match (Figure 6, black curves). We then plotted the

cumulative distributions for each subset as a function of the miRNA-induced change

in expression level (Figure 6). If one type of seed match was highly effective at

causing mRNA expression decreases, we would expect the cumulative distribution of

that subset to be shifted to the left with respect to the black curve. That was the case

for the coding sequence seed matches, although the shift was modest (Figure 6A and

B, green) (P < 0.006 and 10-8

for miR-1 and miR-124, respectively). The 3‟-UTR seed

matches shifted further to the left (Figure 6A and B, red) (P = 0.0004 and 0.00005 for

miR-1 and miR-124 respectively). Thus, coding sequence seed matches appeared to be

effective in decreasing mRNA levels, but 3‟-UTR seed matches were more effective.

These conclusions agree well with previous studies based on other assays and

approaches.

As a further test of the significance of coding sequence seed matches, we compared

how well conserved they were compared to 3‟-UTR seed matches. Of the high

confidence miR-1 and miR-124 targets with 7mer 3‟-UTR seed matches, 41% and

47% of the mRNAs contained seed matches that were perfectly conserved across

mouse, rat and dog. Of the high confidence targets with 7mer coding sequence seed

matches, 25% and 35% of the mRNAs contained conserved seed matches across the

same species. We also compared whether high confidence targets were more likely to

contain conserved seed matches than were non-targets for both coding sequence and

3‟-UTR seed matches. For 7mer 3‟-UTR seed matches, the conservation rate was

higher in Ago2 IP targets than in non-targets (41% vs. 22% for miR-1 and 47% vs.

45% for miR-124; P = 0.005 and 0.3). For 7mer coding sequence seed matches, the

conservation rates were very similar for targets versus non-targets (25% for targets vs.

29% for non-targets for miR-1; 35% for targets vs. 36% for non-targets for miR-124;

P = 0.4 and 0.6). These comparisons strongly suggest that 3‟-UTR seed matches are

more important than coding sequence seed matches for the regulation of mRNA

levels[29,83,88,200,201,202].

65

Figure 6. Comparison of Expression Changes of mRNAs Containing Seed

Matches in 3′-UTRs and Coding Sequences of miR-1 and miR-124 Ago2 IP

Targets.

(A) Cumulative distribution of the change in mRNA levels following transfection with FLAG-Ago2

and miR-1 compared to FLAG-Ago2 alone. This analysis included Ago2 IP targets with 3′-UTR 7mer

seed matches, but no coding sequence 6mer seed matches (21, red), Ago2 IP targets with coding

sequence 7mer seed matches, but no 3′-UTR 6mer seed matches (10, green), and mRNAs that did not

contain 3′-UTR or coding sequence 6mer seed matches (2893, black). Changes in mRNA levels of

Ago2 IP targets with 3′-UTR 7mer seed matches were greater than those for Ago2 IP targets with

coding sequence 7mer seed matches (P = 0.0004), which were in turn greater than those for mRNAs

without any 6mer seed matches in the 3′-UTR or coding sequence (P = 0.006). (B) Same as in (A)

except for mRNAs associated with FLAG-Ago2 upon transfection with miR-124. There were 81 Ago2

IP targets with 3′-UTR 7mer seed matches but no 6mer coding sequence seed matches (red), 43 Ago2

IP targets with coding sequence 7mer seed matches but no 6mer 3′-UTR seed matches (green), and

1877 mRNAs with no 6mer seed matches in their 3′-UTR or coding sequence. Changes in mRNA levels

of Ago2 IP targets with 3′-UTR 7mer seed matches were greater than the changes for Ago2 IP targets

with coding sequence 7mer seed matches (P = 0.0005), which in turn were greater than the changes for

mRNAs without any 6mer seed matches in the 3′-UTR or coding sequence (P<10−8

).

66

Estimation of the Number of mRNAs Regulated by miR-1 and miR-124

Several thousand human mRNAs have seed match sites to either miR-1 or miR-124,

but only a small fraction of these were identified with high confidence as actual

regulatory targets by our gene expression profiling and IP experiments. We estimated

a lower bound of 68 targets for miR-1 and 419 for miR-124, based on the stringent 1%

local FDR criterion. Several prediction algorithms and reporter assays along with our

analysis in Figure 5 suggest that the total number of targets is significantly

higher[83,88,89,197,198,199,209,210]. We therefore took an alternative statistical

approach to this question. We ranked the 7805 (for miR-1-transfected cells) and 7817

(for miR-124-transfected cells) well-measured mRNAs with RefSeq IDs from most

enriched to least enriched in Ago2 IPs (Figure 7A and B, x-axes). Then for groups of

200 sequences, we calculated running averages of the fraction of mRNAs with 7mer

seed matches in their 3‟-UTRs. As expected, the fraction was high for the most

enriched mRNAs and low for the least (Figure 7A and B). We then estimated where

the curve first significantly rose above the background frequency of 7mer seed

matches. To accomplish this we calculated the slope between the right-hand end of the

distribution and every point to the left of it. We took the cutoff to be where the slope

of this line first became negative (Figure 7A and B, vertical gray line). We then

calculated the estimated number of targets as:

number M fM fB

Functions of the High Confidence miR-1 and miR-124 Targets

To determine if miR-1 and miR-124 selectively bind mRNAs that share common

biological functions we searched for enrichment of GO terms in the miRNAs target

sets identified through Ago2 immunopurification. There is modest enrichment of

several GO categories for each miRNA target set: for example, the miR-124 set is

enriched for mRNAs encoding proteins localized to the membrane (P = 0.002) or that

bind GTP (P = 0.009), and the miR-1 set is

67

Figure 7. Estimation of the Number of miR-1 and miR-124 Targets.

(A) Moving average plot (window size of 200) of the fraction of mRNAs with 7mer 3′-UTR seed

matches to miR-1. mRNAs were ranked by their SAM enrichment in Ago2 IPs in the presence of miR-1

compared to Ago2 alone, with 1 corresponding to the most enriched mRNA. To determine the point at

which the curve first rose above the background level of 7mer seed matches, we first calculated the

slope of each least-squares-fit regression line between the right-hand end of the distribution and every

point to left of it. The point at which the curve first rose above the background level of 7mer seed

matches was determined as the point at which the slope was first negative (vertical grey line). The

fraction of mRNAs containing 7mer seed matches to the right of the vertical grey line was considered to

be the background level of 7mer seed matches (horizontal grey line). To estimate the total number of

targets (pink shaded region), the number of mRNAs to the left of the vertical grey line (3071 of 7805)

was multiplied by the fraction of mRNAs to the left of the vertical line containing 7mer 3′-UTR seed

matches (0.23) minus the fraction of mRNAs to the right of the vertical line containing 7mer seed

matches (0.12). This results in an estimate of 325 targets. (B) Same as in (A), but for miR-124. 6393 of

7817 mRNAs were to the left of the vertical grey line. The fraction of mRNAs with 7mer 3′-UTR seed

matches to the left of the grey vertical line was 0.21, while the fraction of mRNAs with 7mer seed

68

matches to the right of the grey vertical lines was 0.07. This results in an estimate of 1000 targets. (C)

Same as in (A), except moving average plots of the fraction of mRNAs with 7mer coding sequence seed

matches to miR-1. mRNAs with 6mer 3′-UTR seed matches were removed, leaving 4855 mRNAs. 820

mRNAs were to the left of the vertical grey line. The fraction of mRNAs with 7mer seed matches to the

left and right of the grey vertical line was 0.24 and 0.14 respectively. This results in an estimate of 83

targets. (D) Same as in (C), but for miR-124. 3312 of 3916 mRNAs were to the left of the vertical grey

line. The fraction of mRNAs with 7mer seed matches to the left and right of the grey vertical line was

0.15 and 0.08 respectively. This results in an estimate of 236 targets.

where M is the number of mRNAs to the left of the cutoff, fM is the fraction of those mRNAs with 7mer

seed matches, and fB is the fraction of the mRNAs to the right of the cutoff with 7mer seed matches.

This method gives estimates of 325 and 1000 mRNAs recruited to Ago2 by 3‟-UTR seed matches to

miR-1 and miR-124, respectively (Figure 7A,B). Using 6mer seed matches rather than 7mers yielded

similar estimates, 293 and 1232 respectively (data not shown). Applying the same logic to mRNAs with

7mer seed matches in their coding sequences but no 6mer 3‟-UTR seed matches gives estimates of 83

miR-1 and 236 miR-124 targets recruited to Ago2 exclusively through miRNA targeting of the coding

sequence (Figure 7C and D). Using 6mer seed matches again yielded similar estimates (50 and 253;

data not shown). These data provide direct empirical evidence that miR-1 and miR-124 have hundreds

of direct mRNA targets. These miRNAs are highly connected hubs in the network of RNA regulation.

enriched for mRNAs involved mRNA metabolism (P = 0.006) and cell motility (P =

0.015). We also tested the distribution of ~430 curated gene sets in the IP enrichments

as a whole. Using curated gene sets from gene set enrichment analysis[261] rather

than GO terms, we found that no gene sets were significantly enriched at a corrected

P-value threshold of 0.05.

Using Ago2 Immunopurification Enrichment and mRNA Expression Changes to

Assess Computational Target Prediction Methods

Our empirical data on miR-1 and miR-124 targets allow us to assess computational

methods for the prediction of miRNA targets. We examined five methods: TargetScan

4.0, which looks for seed matches in appropriate sequence contexts[88]; TargetScan

3.0 and PicTar, which look for seed matches conserved among human, dog, mouse,

rat, and chicken mRNAs[89,199]; PITA, which makes use of seed matches and

predicted target accessibility[209]; and MiRanda, which looks at sequence

complementarity and conservation among human, mouse and rat mRNAs[197,198].

69

The performance of these methods was assessed by cumulative distribution plots, with

either microarray expression data or Ago2 IP enrichment data on the x-axis (Figure

S3). The more successful the method, the further its cumulative distribution curve

should shift to the left (for expression data) or the right (for Ago2 IP enrichment). As

shown in Figure S3, TargetScan 4.0 performed best for predicting both miR-1 targets

and miR-124 targets (blue curves). TargetScan 3.0 and PicTar were next best

(magenta and orange curves), followed by PITA and Miranda (green and gray curves).

None of the computational methods performed as well as an expression data-plus-seed

match criterion for predicting Ago2 IP-enriched targets (Figure S3A and B, red

curves). Likewise, none of the computational methods performed as well as an Ago2

IP-enrichment-plus-seed match criterion for predicting targets identified by the

expression data (Figure S3C and D, red curves). Thus, while TargetScan 4.0

performed particularly well, the two empirical methods (expression data and Ago2 IP

enrichment) were superior. This indicates that some information important for miRNA

target recognition is still missing from the prediction algorithms.

70

Discussion

A Direct Assay to Identify Targets of Specific miRNAs

Much of what we know about miRNA targeting has been inferred indirectly from

effects on mRNA levels, from phylogenetic conservation of recognition sites, and, on

a smaller scale, from effects on levels of the encoded proteins. Such studies have

provided a foundation for our understanding of miRNA regulation of gene expression,

yet crucial information directly linking these effects to miRNA pathways has been

missing. We would like to know what mRNAs are recruited to RISC by each miRNA

and are thus acted upon by miRNA-mediated regulation.

A simple method employing immunoaffinity isolation of Ago2, a core component of

RISC, identifies mRNAs recruited to RISC by specific miRNAs. Knowledge of these

mRNAs provides a direct and critical point of reference for understanding the

molecular mechanism and logic of mRNA target specificity and for comprehensive

investigation of the functional consequences of miRNA-induced interactions.

The selective association of specific mRNAs with Ago2 in response to specific

miRNAs is prima facie evidence for their miRNA-mediated recruitment to

Ago2/RISC. The enrichment of cognate seed matches in computationally predicted

favorable contexts and the correlation between specific miRNA-dependent Ago2-IP

enrichment and changes in mRNA levels are strong evidence that the assay reflects a

direct and functional interaction between the transfected miRNA and the Ago2 IP-

enriched mRNAs. Although the introduction of exogenous FLAG-Ago2 could have

altered the normal specificity of these interactions, the negligible effect of exogenous

FLAG-Ago2 on global patterns of expression of either mRNAs or miRNAs argue

against a major distortion of normal regulation and suggest that the interactions we

observed are generally faithful representations of native interactions.

71

The evolutionary conservation of many of the seed matches to miR-1 and miR-124 in

the 3‟-UTRs of the mRNAs identified as specific targets in our assay lends further

credence to the biological relevance of the interactions. Transfection of miR-1 or miR-

124 into these cells greatly increased the cellular levels of these miRNAs (which are

undetectable in untransfected cells), but the resulting concentrations appeared to be

well below those of the most highly expressed endogenous miRNAs, based on qRT-

PCR experiments (data not shown). The lack of enrichment (or underenrichment) of

seed matches to other miRNAs in the mRNAs recruited to Ago2-RISC after

transfection of miR-1 or miR-124 further implies that perturbation of endogenous

miRNAs and miRNA targets was not a significant factor in the Ago2-enriched

mRNAs.

The cells we used in this study were chosen for their experimental tractability; they

were not an optimal model for studies of the regulatory roles of miR-1 or miR-124.

The Ago2 IP procedure used herein should, however, be widely applicable to other

cells or even whole organisms, in which mRNAs identified as targets, by either

overexpressing or blocking a specific miRNA, can be related to specific biological

consequences.

Although there were significant changes in the levels of many of the mRNAs recruited

to Ago2/RISC in the presence of specific miRNAs, there were also many associated

mRNAs that were only slightly altered in expression level. The ability to identify

mRNA targets directly, without relying on a change in their levels in response to

perturbation of a specific miRNA, makes it possible to systematically investigate other

possible miRNA-directed effects on their expression, including, for example, effects

on subcellular localization or translation.

Functional Insights into miRNA Targeting and Regulation

The strong correlation between miRNA-specific association with Ago2 and decreases

in mRNA levels for mRNA targets with 3‟-UTR seed matches suggests that the

strength or properties of a miRNA‟s association with a potential target mRNA has an

72

important role in regulating their degradation. Most high confidence mRNA targets

with 3‟-UTR seed matches were regulated to some degree at the level of mRNA

abundance, albeit weakly in most cases.

Comparing the magnitude of the changes in mRNA abundance to direct measurements

of the efficiency with which each mRNA is recruited to RISC revealed a quantitative

relationship between the recruitment efficiency and consequences for expression.

Moreover, the link between size, number of seed matches and recruitment efficiency

suggests an evolutionary mechanism for quantitative tuning of the regulatory response

of each mRNA to a miRNA. A continuous scale of regulation, tuning affinity, or

context of a site that allows an increase in existing regulation is easier than evolving a

functional site de novo.

Insights into miRNA-based Regulation From Recent, Related Publications

While this paper was in preparation, four papers describing similar strategies to

identify mRNAs associated with RISC and specific miRNAs were

published[200,262,263,264]. We briefly review these results, highlighting the

similarities that strengthen common conclusions and differences that may be

instructive with respect to methodologies and biological mechanisms.

Easow et al. [200] immunopurified affinity-tagged Ago1 from Drosophila

melanogaster S2 cells and identified associated mRNAs via microarray hybridization.

The authors found 89 mRNAs specifically associated with Ago1. 3‟-UTR seed

matches to some highly expressed miRNAs were overrepresented in the target set.

The authors also found some enrichment of coding sequence seed matches to highly

expressed miRNAs in the Ago1 targets. The efficacy of coding seed match sites was

tested for two mRNAs lacking seed matches in their 3‟-UTR by cloning the coding

sequences in-frame with a luciferase reporter; mutation of these seed match sites led to

a ~25% increase in the expression of luciferase. The authors also created two

exogenous seed matches to a highly expressed miRNA in the coding sequence of

luciferase by mutating silent codon positions. Sequences with these seed matches

73

were also introduced into the SV40 3‟-UTR downstream of a luciferase reporter.

Addition of the seed match sites to both regions led to decreased luciferase levels, but

the repression was more pronounced when the sites were in the 3‟-UTR. 108 mRNAs

were differentially underenriched in Ago1 IPs of embryos with a mutation in miR-1

compared to Ago1 IPs from lysates of embryos with the wild-type miR-1 gene. The

regulatory effect of miR-1 on potential targets was gauged by comparing the luciferase

levels of reporter genes containing 3‟-UTRs of 11 of the 32 potential miR-1 targets in

the presence or absence of miR-1. Luciferase activity was reduced in the presence of

miR-1 for all 11 constructs, but to different extents, varying from ~5-60%. In all

cases, the decreases in luciferase activity were greater than the decreases in mRNA

levels, consistent with substantial translational regulation, as suggested by numerous

studies.

Zhang et al. [264] immunopurified GFP-tagged AIN-1 and AIN-2, which they found

to be Ago-associated proteins, from Caenorhabditis elegans whole animals and

identified associated miRNAs by sequencing and associated mRNAs by DNA

microarray hybridization. The authors found approximately 3000 (15% of all known

C. elegans genes) mRNAs associated with either AIN-1 or AIN-2, including many

known and predicted miRNA targets.

Beitzinger et al. [262] immunopurified Ago1 and Ago2 from HEK293T cells and

identified some associated mRNAs by sequencing. About 600 clones derived from

RNAs associated with Ago1 or Ago2 were sequenced. Nonspecific interactions with

the resin or antibody were not controlled for by comparison to “mock”

immunopurification. Instead, clones recovered once were classified as nonspecific and

clones recovered multiple times were counted as direct targets. Using this criterion,

only 82 unique Ago1 and 28 unique Ago2 targets were found, with 15 targets in

common. The large number of single hit clones indicates that the results represent a

non-exhaustive list of Ago2 targets. Thus, it is not surprising that the Ago2-associated

mRNAs identified in this study do not significantly overlap with our findings.

74

Karginov et al. [263] applied an Ago2 immunopurification strategy similar to ours to

identify Aog2-associated mRNAs and miRNAs in HEK293T cells. The authors found

over one thousand mRNAs specifically associated with Ago2. Ago2 IP enrichment

was positively correlated with the presence of 3‟-UTR seed match sites to highly

expressed miRNAs. Following transfection of miR-124, 370 mRNAs were specifically

recruited to Ago2. About half of the putative targets contained 7merm-8 seed matches

to miR-124 in their 3‟-UTRs. The mRNA levels of many of the putative miR-124

targets were significantly decreased in response to the presence of miR-124. The 3‟-

UTRs from 21 of 30 mRNAs that were miR-124 IP targets but did not change

significantly at the mRNA level that were incorporated into luciferase probes lead to

significant miR-124 dependent decreases in protein expression. Of the 370 mRNAs

for which Karginov et al. reported a miR-124 dependent association with Ago2, 238

had unique RefSeq IDs and were detected on our arrays. Forty-nine percent of these

mRNAs were also classified as miR-124 targets in our experiments at a 1% local FDR

threshold, 59% were enriched at a less stringent 10% local FDR cut-off, and 95% were

more enriched than the median IP enrichment of all mRNAs. Our findings are thus in

broad general agreement, and provide further validation of the approach and

individual insights into miRNA-based regulation.

There were also findings that were unique to our study. Enrichment of mRNAs

containing coding sequence and 5‟-UTR seed match sites was not observed in the

Karginov et al. study. The discrepancies may be related to differences in the IP

procedures. Karginov et al. washed immunopurified beads with 650 mM NaCl,

whereas our immunopurifications and washes were performed with 150 mM KCl. In

our hands, washing the Ago2 immunopurified beads with 300 mM KCl resulted in

loss of enrichment of total RNA and Dicer, whereas Ago2 enrichment was not

affected (SFigure 1). We did not analyze the pool of mRNA that remained bound to

Ago2 following this more stringent wash. It is possible that the stringent wash

employed in Karginov et al. disrupted relatively labile mRNA:RISC interactions,

including those in coding sequences and 5‟-UTRs. On the other hand, it is possible

that coding sequence and 5‟-UTR interactions identified in our assay were “created”

75

by association with Ago2 post-lysis. Another reason for the discrepancy could be

differences in the growth rate of the cells at the time of lysis. The global ribosome

occupancy at the time of lysis in the cells used in our experiments was quite low

because the cells reached high confluency during the 48 hours between transfection

and cell-lysis. It is possible that under conditions in which there is more translation,

coding sequence and 5‟-UTR sites are relatively less occupied compared to 3‟-UTR

sites, because these miRNA-mRNA interactions are disrupted by ribosomes. The

Karginov et al. study employed similar growth conditions, and also prepared extracts

48 hours after transfections. Regardless of the causes for the discrepancies, the

significant decrease in abundance of mRNAs containing coding sequence sites

suggests that these are biologically relevant miRNA targets.

76

Materials and Methods

Plasmids and oligonucleotides

CMV-FLAG-Ago2 plasmid was provided by G. Meister and T. Tuschl[63].

miR-1 siRNA:

sense: 5‟UGGAAUGUAAAGAAGUAUGUA3‟

antisense: 5‟CAUACUUCUUUACAUUCAAUA3‟

miR-124 siRNA:

sense: 5‟UAAGGCACGCGGUGAAUGCCA3‟

antisense: 5‟GCAUUCACCGCGUGCCUUAAU3‟

Cell culture and transfection

HEK293T cells were obtained from ATCC (Cat #CRL-11268) and grown in

Dulbecco‟s modified Eagle‟s medium (Invitrogen) with 10% fetal bovine serum

(Invitrogen) and supplemented with 100U/ml penicillin, 100mg/ml streptomycin, and

additional 4mM glutamine (Invitrogen) at 37°C and 5% CO2. Transfections of

HEK293T cells were carried out with calcium phosphate. Cells were plated in 10cm

dishes 12-16hrs prior to transfection at 30% confluency. For 1mL transfection

mixtures (1/10 volume of growth media) 61µl of 2M CaCl2 and 10µg of Ago2 plasmid

DNA were diluted into 500 µl of nuclease free H2O (Invitrogen) and added slowly to

500 µl of 2X HBS (50mM Hepes, 280mM NaCl, 1.5mM Na2HPO4) pH 7.1. After ~1

minute, the mixture was added to a 10cm plate at a medium pace. Transfections with

the miR-1 and miR-124 oligonucleotides were performed analogously by diluting a

40µM stock to 5nM in the 500µl nucleic acid mixture along with the plasmid DNA.

Imunoaffinity purification and RNA isolation

For each purification, 400µl of 4°C lysis buffer (150mM KCl, 25mM Tris-HCl pH

7.4, 5mM EDTA, 0.5% Nonidet P-

SUPERase•In

77

(Ambion Cat #AM2694) was added to ten 10cm plates after washing 1X in PBS 48

hrs post-transfection. After 30min at 4°C, the plates were scraped and the lysates

combined and spun at 4°C for 30 minutes at 14,000 RPM in a microcentrifuge. The

supernatant was then collected and filtered through a 0.45µm syringe filter. The lysate

was then mixed with 300µl of FLAG resin (Sigma Cat #A2220), which was

equilibrated by washing 2X with lysis buffer with 10X volume. The beads were

incubated with the lysate for 4hrs at 4°C and washed 2X with 10X volume of lysis

buffer for 5 minutes. Five percent of the beads were frozen for SDS PAGE analysis

after the second wash. RNA was extracted directly from the remaining beads with

25:24:1 phenol:chloroform:isoamyl alcohol (Invitrogen Cat#15593-031). Trace

amounts of phenol were removed by chloroform extraction and RNA was precipitated

using sodium acetate with Glyco-Blue (Ambion Cat# AM9516) as a carrier. RNA

pellets were resuspended in 25µl of RNase free water and stored at -80 °C. Small

RNA samples for PAGE detection were isolated using a modified protocol for RNA

isolation using Invitrogen‟s Micro-to-Midi kit (Invitrogen Cat#12183-18)[139]. Small

RNA for microarray analysis was fractionated using the FLASH-PAGE system

(Ambion Cat#AM13100) as per vendors‟ instructions.

Western blots, Sypro Staining, and Nucleic Acid PAGE

Resin saved from each immunoaffinity purification were resuspended in water and

diluted to 1X sample buffer and 1X reducing buffer (Biorad Cat#161-0791, and 161-0792)

and heated at 95oC for 3 min. Each sample was then divided and one-half was loaded

onto either a 4-12% criterion XT gel (BioRad Cat# 345-126) for protein staining with

sypro ruby (Invitrogen Cat #S-12000) or onto a 3-8% criterion XT gel for western

blotting (BioRad Cat# 345-0129). For the SYPRO ruby staining, gels were treated as

per the vendor‟s instruction immediately after electrophoresis. For western analysis,

each gel was transferred onto polyvinylidene fluoride membrane (Immobilon Cat#

IPVH08100) and probed with FLAG m2 antibody (Sigma Cat# F-1804) and a

polyclonal Dicer antibody generated by rabbits inoculated with a peptide

corresponding with the N-terminus of Dicer: EILRKYKPYERQQFESVC (Quality

Controlled Biochemicals). Small RNA was detected using 15% urea TBE criterion

78

gels (Biorad Cat# 345-0055) and Syber Gold (Invitrogen Cat# S-1149) as per the

vendor‟s instructions. RNA (0.5-1 µg) was loaded in each lane.

Microarray Production and Pre-hybridization Processing

Detailed methods for microarray experiments are available at the Brown lab website

(http://cmgm.stanford.edu/pbrown/protocols/index.html). HEEBO oligonucleotide

microarrays and miRNA microarrays were produced by Stanford Functional Genomic

Facility. The HEEBO microarrays contain ~45,000 70-mer oligonucleotide probes,

representing ~30,000 unique genes. A detailed description of this probe set can be

found at (http://microarray.org/sfgf/heebo.do)[255]. The miRNA arrays (Ambion

miRNA Bioarrays version 2)[256] contained probes for 668 human, mouse and rat

miRNAs. Each probe was printed in duplicate.

RNA from immunopurification experiments was hybridized to microarrays printed on

aminosilane-coated glass (Schott Nexterion A). Prior to hybridization, the

oligonucleotides were cross-linked to the aminosilane-coated surface with 65mJ of

UV irradiation. Slides were then incubated in a 500 ml solution containing 5X SSC

(1X SSC = 150 mM NaCl, 15 mM sodium citrate, pH 7.0), 1% w/v Blocking Reagent

(Roche Cat# 1109617001), and 0.1% SDS for 35 minutes at 65oC. Slides were washed

twice for 1 min each in glass chambers containing 400 ml water, dunked in a glass jar

containing 400 ml 95% ethanol for 15 seconds, then dried by centrifugation. Slides

were used the same day.

mRNA expression experiments and miRNA experiments were performed with

microarrays printed on epoxysilane-coated glass (Schott Nexterion E). Prior to

hybridization, slides were first incubated in a humidity chamber (Sigma Cat# H6644)

containing 0.5X SSC for 30 min at room temperature. Slides were snap-dried at 70-

80oC on an inverted heat block. The free epoxysilane groups were blocked by

incubation with 1M Tris-HCl pH 9.0, 100 mM ethanolamine (Sigma Cat# E9508), and

0.1% SDS for 20 minutes at 50oC. Slides were washed twice for 1 min each with 400

ml water, and then dried by centrifugation. Slides were used the same day.

79

Sample Preparation, Hybridization and Washing

For HEEBO microarray experiments, poly-adenylated RNAs were amplified in the

presence of aminoallyl-UTP with Amino Allyl MessageAmp II aRNA kit (Ambion

Cat# 1753). For expression experiments, universal reference RNA was used as an

internal standard to enable reliable comparison of relative transcript levels in multiple

samples (Stratagene Cat# 740000). Amplified RNA (3- ntly

labeled with NHS-monoester Cy5 or Cy3 (GE HealthSciences Cat# RPN5661). Dye-

solution containing 3X SSC, 25mM Hepes- -1 DNA

(Invitrogen Cat# 152790

tRNA (Invitrogen Cat # 15401029), and 0.3% SDS. The sample was incubated at 70oC

for 5 minutes, spun at 14,000 rpm for 10 minutes in a microcentrifuge, then hybridized

at 65oC for 12-16 hours. For immunopurification experiments, microarrays were

hybridized inside sealed chambers in a water bath using the M-series lifterslip to

contain the probe on the microarray (Erie Scientific Cat # 22x60I-M-5522). For

mRNA expression experiments, microarrays were hybridized using the MAUI

hybridization system (BioMicro), which promotes active mixing during hybridization.

Following hybridization, microarrays were washed in a series of four solutions

containing 400 ml of 2X SSC with 0.05% SDS, 2X SSC, 1X SSC, and 0.2X SSC,

respectively. The first wash was performed for 5 minutes at 65oC. The subsequent

washes were performed at room temperatures for 2 minutes each. Following the last

wash the microarrays were dried by centrifugation in a low-ozone environment (<5

ppb) to prevent destruction of Cy dyes[265]. Once dry, the microarrays were kept in a

low-ozone environment during storage and scanning (see

http://cmgm.stanford.edu/pbrown/protocols/index.html).

Small RNAs for miRNA microarrays were labeled using the mirVana labeling kit

(Ambion Cat# Am1562) and samples were prepared, hybridized and washed

according to manufacturer‟s instructions.

80

Scanning and Data Processing

Microarrays were scanned using either AxonScanner 4200 or 400oB (Molecular

Devices). PMT levels were auto-adjusted to achieve 0.1-0.25% pixel saturation. Each

element was located and analyzed using GenePix Pro 5.0 (Molecular Devices). These

data were submitted to the Stanford Microarray Database for further analysis

(http://smd.stanford.edu/cgi-bin/publication/viewPublication.pl?pub_no=685)[266].

Data were filtered to exclude elements that did not have a regression correlation of

≥0.6 between Cy5 and Cy3 signal over the pixels compromising the array element of

and intensity/background ratio of ≥2.5 in at least one channel, for 60% of the arrays.

For cluster and SAM (Fig2) analysis of Ago2 +/- miR-1/124 IPs versus mock IPs,

measurements corresponding to oligonucleotides that map to the same entrezID were

treated separately and the data were globally normalized per array, such that the

median log2 ratio was 0 after normalization. For all analysis after Figure 3,

measurements corresponding to oligonucleotides that map to the same entrezID were

averaged and the data were globally normalized per array, such that the median log2

ratio was 0 after normalization. To control for variation among groups of experiments

performed at different times, each group was normalized by subtracting the median

log2 ratio for each gene across the experiments in a group from the log2 ratio of the

gene in each experiment. The groups are labeled in the supplementary information.

miRNA microarray experiments were normalized by subtracting the median value of

average Cy3 and Cy5 signal intensities of negative control spots from the average Cy3

and Cy5 signal of each experimental measurement. Normalized Cy3 and Cy5 signal

intensities from replicate experiments were normalized and log2 transformed

(measurements with negative values were changed to a value of 1). The distribution of

the log2 signal intensities was nearly bimodal; miRNAs with signal intensity greater

than the value at the trough of the distribution were considered to be expressed.

The microarray data have been submitted to Gene Expression Omnibus (GEO)

(www.ncbi.nlm.nih.gov/geo/) under the accession number GSE11082.

81

Microarray Analyses

Hierarchical clustering was performed with Cluster 3.0[267] and visualized with Java

TreeView 1.0.12[268].

For SAM, unpaired two-class t-tests were performed with default settings (R-package

samr). FDRs were generated from up to 1000 permutations of batch normalized (see

above) data.

Sequence Data

For each entrezID, the RefSeq sequence with the longest 3‟-UTR was used. In cases

where there were multiple RefSeqs with the same 3‟-UTR length, the one that was

alpha-numerically first was used. RefSeq 3‟-UTR, coding, and 5‟-UTR sequences

were retrieved from UCSC genome browser (hg18). Seed match sites in these

sequences were identified with Perl scripts. miR-1 seed matches: 6mer_n2-7

“CAUUCC”, 6mer_n3-8 “ACAUUC”, 7mer-m8 “ACAUUCC”, 7mer-A1

“CAUUCCA”, 8mer “GUGCCUUA”. miR-124 seed matches: 6mer_n2-7

“UGCCUU”, 6mer_n3-8 “GUGCCU”, 7mer-m8 “GUGCCUU”, 7mer-A1

“UGCCUUA”, 8mer “GUGCCUUA”.

Conservation of Seed Match Sites

For each RefSeq, the 28-way multiple sequence alignments files for the 3‟-UTR or

coding sequences were retrieved from UCSC genome table browser. The human, dog

(canFam2), mouse (mm8), and rat (rn4) sequences were extracted and multiple

sequence alignments files corresponding to the same RefSeq were stitched together

with Galaxy. Sites with 7mer-m8 or 7mer-A1 matches present in all three species

within 20 positions of the human seed match site were considered to be conserved.

Sequence Analyses

Enrichment of hexamers in 3‟-UTRs of miR-1 and miR-124 targets relative to

nontargets was performed on the Regulatory Sequence Analysis Tools website. P-

82

values were calculated with binomial distribution function and corrected for multiple

hypothesis testing using the bonferroni method.

To identify sequence motifs associated with enrichment in immunopurifications

subsequent to miR-1 and miR-124 transfection, the MEME method was used[260].

MEME 3.0.14 was downloaded from the MEME website and run with default settings,

except searching the forward strand only with zoops model.

miRNA Target Predictions

Predictions for Targetscan 4.0 were downloaded on June 12, 2007. Context scores for

each miRNA site in each RefSeq sequence were summed to get the cumulative context

score for that miRNA. Predictions for human PITA3/15 flank were downloaded on

November 4, 2007. For both TargetScan 4.0 and PITA, miR-506/124-2 predictions

were used for miR-124, because of changes in annotated sequences in Sanger miRNA

database. These miRNAs share the same 5‟end as the miR-124 sequence used in this

study, but have different 3‟ends. Predictions for TargetScan 3.0 were retrieved on

November 2, 2006. Predictions for Pictar 5-way and Miranda were retrieved on March

2, 2007. Miranda predictions used Ensemble IDs, which were mapped to RefSeqs

using UCSC genome table browser.

Gene Ontology and Gene-set Analyses

Enrichment of GO terms in miR-1, miR-124, and Ago2 target sets was identified with

Expression Analysis Systematic Explorer[269]. Enrichment of gene sets was

performed with Gene Set Enrichment Analysis[261].

Acknowledgements

Drs. Tom Tuschl and Günther Meister provided the FLAG-Ago2 plasmid, and we

thank Drs. Tongbin Li and Jason Myers and members of the Brown, Ferrell, and

Herschlag labs for discussions and critical reading of the manuscript.

83

Supplementary Figures

Figure S1. Disassociation of Dicer from Ago2 IPs in 300 mM KCL

Disassociation of Dicer from Ago2 IPs in 300 mM KCL. Western blot with Dicer antibody on protein

associated with an Ago2 IP from cells lysed in 150 mM KCl (left), after washing once with 300 mM

KCL (middle), and after washing a second time with a 300 mM KCl concentration (right).

84

Figure S2. The Length and Number of 3′-UTR Seed Match Sites to miR-1 and

miR-124 Correlates With Enrichment in Ago2 IPs.

(A) Cumulative distributions of Ago2 IP enrichment due to the presence of miR-1 of mRNAs

containing single 6–8mer 3′-UTR seed match sites. The IP enrichment of mRNAs containing different

6–8mer sites was as follows: 8mer (144, red)>7mer-m8 (257, green, P = 0.005, one-sided Mann–

Whitney test)>7mer-A1 (375, blue, P = 0.0005)>6mer_n-7 (magenta, 567, P = 0.005)~6mer_n3–8

(orange, 654, P = 0.3)~no seed match (black, 4855, P = 0.2). (B) Same as in (A), except for miR-124.

The IP enrichment of mRNAs containing different 6–8mer sites was as follows: 8mer (78, red)>7mer-

m8 (329, green, P<10-4)>7mer-A1 (283, blue, P<10-4)>6mer_n-7 (magenta, 845, P<10-9)~6mer_n3–8

(orange, 624, P = 0.09)>no seed match (black, 3916, P<10-4). (C) CDF of Ago2 IP enrichment due the

presence of miR-1 of mRNAs containing multiple 7mer 3′-UTR seed match sites (61, red), single 7mer

3′-UTR seed match sites (632, green), and no 3′-UTR seed match sites (black). mRNAs with multiple

miR-1 7mer 3′-UTR seed match sites were significantly more enriched than mRNAs containing single

7mer seed match sites (P = 0.03). (D) Same as in (C), except for miR-124. mRNAs containing multiple

7mer 3′-UTR sites (81) were more enriched than mRNAs with single 7mer sites (612, P = 0.01).

85

Figure S3. Using Ago2 IP Enrichment and mRNA Expression Changes to Assess

Computational Target Prediction Methods.

(A) Cumulative distributions of Ago2 IP enrichment in cells transfected with miR-1 relative to cells

transfected with FLAG-Ago2 alone of several sets of mRNAs predicted to be targeted by miR-1. The

100 mRNAs containing 7mer 3′-UTR seed matches to miR-1 whose levels were most significantly

downregulated due to the presence of miR-1 were the most enriched group (red). The 100 mRNAs

whose levels were most significantly downregulated due to the presence of miR-1 irrespective of seed

match sites were the next most enriched group in Ago2 IPs (dark green). The 100 mRNAs containing

the lowest TargetScan 4.0 “cumulative context score” were the next most enriched group (blue).

TargetScan 3.0 (274, magenta) and PicTar 5way predictions (97, orange) were the next most enriched

groups. The 100 mRNAs containing the most favorable 3′-UTRs for miR-1 binding according to PITA

3/15 flank (green), miRanda predictions (113, grey) and mRNAs containing 7mer 3′-UTR seed matches

86

were the next most enriched groups (1245, cyan). mRNAs containing no 6mer 3′-UTR seed matches

were the least enriched group (4765, black). (B) Same as in (A) except for miR-124. 573 mRNAs were

TargetScan 3.0 predictions; 128 mRNAs were PicTar 5way predictions, 202 mRNAs were miRanda

predictions, 1500 mRNAs contained 7mer 3′-UTR seed matches, and 3820 mRNAs did not contain a

6mer 3′-UTR seed match. (C) Cumulative distributions of changes in mRNA levels in cells transfected

with miR-1 relative to cells transfected with FLAG-Ago2 alone of several sets of mRNAs predicted to

be targeted by miR-1. The 100 mRNAs containing 7mer 3′-UTR seed matches to miR-1 that were most

enriched in Ago2 IPs due to the presence of miR-1 was the most underenriched group (red). The 100

mRNAs that were most enriched in Ago2 IPs due to the presence of miR-1 irrespective of seed match

sites was the next group (dark green). The 100 mRNAs containing the lowest TargetScan 4.0

“cumulative context score” was the next most underenriched group (blue). TargetScan 3.0 (magenta)

and PicTar 5way predictions (orange) were the next most underenriched groups. The 100 mRNAs

containing the most favorable 3′-UTRs for miR-1 binding according to PITA 3/15 flank (green),

miRanda predictions (grey) and mRNAs containing 7mer 3′-UTR seed matches were the next most

underenriched groups (cyan). mRNAs containing no 6mer 3′-UTR seed match was the least

underenriched group (black). (D) Same as in (C) except for miR-124.

87

Concordant Regulation of Translation and mRNA

Abundance for Hundreds of Targets of a Human microRNA

David G. Hendrickson1, Daniel J. Hogan

2,3, Heather L. McCullough2,3,4, Jason W.

Myers2,3, Daniel Herschlag

2*, James E. Ferrell

1,2*, Patrick O. Brown

2,3*

1 Department of Chemical and Systems Biology, Stanford University School of

Medicine, Stanford, California, United States of America, 2 Department of

Biochemistry, Stanford University School of Medicine, Stanford, California, United

States of America, 3 Howard Hughes Medical Institute, Stanford University School of

Medicine, Stanford, California, United States of America, 4 Department of Genetics,

Stanford University School of Medicine, Stanford, California, United States of

America

These authors contributed equally to this work.

This chapter was reprinted from:

PLoS Biology 2009 Nov;7(11):e1000238

PLoS journals publish under the Creative Commons Attribution License (CCAL).

No permission is required from the authors or the publishers.

DGH, DJH, JWM and POB conceived and designed experiments. DGH, DJH, and

JWM performed the experiments, analyzed the data, and helped write the manuscript.

POB, JEF, and DH provided general direction, helped with the data analysis and

figure reparation, and helped write the manuscript.HLM provided reagents.

88

Abstract

MicroRNAs (miRNA) regulate gene expression posttranscriptionally by interfering

with a target mRNA‟s translation, stability, or both. We sought to dissect the

respective contributions of translational inhibition and mRNA decay to microRNA

regulation. We identified direct targets of a specific miRNA, miR-124, by virtue of

their association with Argonaute proteins, core components of miRNA effector

complexes, in response to miR-124 transfection in human tissue culture cells. In

parallel, we assessed mRNA levels and obtained translation profiles using a novel

global approach to analyze polysomes separated on sucrose gradients. Analysis of

translation profiles for ~8,000 genes in these proliferative human cells revealed that

basic features of translation are similar to those previously observed in rapidly

growing Saccharomyces cerevisiae. For ~600 mRNAs specifically recruited to

Argonaute proteins by miR-124, we found reductions in both the mRNA abundance

and inferred translation rate spanning a large dynamic range. The changes in mRNA

levels of these miR-124 targets were larger than the changes in translation, with

average decreases of 35% and 12%, respectively. Further, there was no identifiable

subgroup of mRNA targets for which the translational response was dominant. Both

ribosome occupancy (the fraction of a given gene‟s transcripts associated with

ribosomes) and ribosome density (the average number of ribosomes bound per unit

length of coding sequence were selectively reduced for hundreds of miR-124 targets

by the presence of miR-124. Changes in protein abundance inferred from the observed

changes in mRNA abundance and translation profiles closely matched changes

directly determined by Western analysis for 11 of 12 proteins, suggesting that our

assays captured most of miR-124–mediated regulation. These results suggest that

miRNAs inhibit translation initiation or stimulate ribosome drop-off preferentially

near the start site and are not consistent with inhibition of polypeptide elongation, or

nascent polypeptide degradation contributing significantly to miRNA-mediated

regulation in proliferating HEK293T cells. The observation of concordant changes in

mRNA abundance and translational rate for hundreds of miR-124 targets is consistent

89

with a functional link between these two regulatory outcomes of miRNA targeting,

and the well-documented interrelationship between translation and mRNA decay.

Introduction

MicroRNAs (miRNAs) are small noncoding RNAs whose complementary pairing to

target mRNAs potentially regulates expression of more than 60% of genes in many

and perhaps all metazoans [9,27,59,83,84,85]. Destabilization of mRNA and

translational repression have been suggested as the mechanisms of action for miRNAs

[9,25,26,27,28,29,30,31,33,86,106], and recent work directly measuring endogenous

protein levels in response to altered miRNA expression levels found that specific

miRNAs modestly inhibit the production of hundreds of proteins [90,91].

The importance and functional range of miRNAs are evidenced by the diverse and

often dramatic phenotypic consequences when miRNAs are mutated or misexpressed,

leading to aberrant development or disease [30,92,93,94,95,96,97,98]. Although

regulation by miRNAs is an integral component of the global gene expression

program, there is currently no consensus on either the mechanism by which they

decrease the levels of the targeted proteins or even the steps in gene expression

regulated by miRNAs [27,32,99,110,111,112].

The proposal that miRNAs decrease protein levels without affecting mRNA stability

arose from the observation that the miRNA lin-4 down-regulates lin-14 expression in

the absence of noticeable changes in lin-14 mRNA abundance in Caenorhabditis

elegans [8,12,30,270,271]. Subsequent studies in mammalian cell culture provided

further support for this model [58,60,217,218]. Several studies have found that

repressed mRNAs as well as protein components of the miRNA regulatory system

accumulate in P-bodies, suggesting that repressed mRNAs may be sequestered away

from the translation pool [56,108,194,195,214,215,216]. Other evidence points to

deadenylation of miRNA-targeted mRNAs, an effect that can inhibit translation

[33,57,61,102,237,272,273,274,275,276]. Some studies have argued that initiation of

90

translation is blocked at either an early, cap-dependent stage or later during AUG

recognition or 60S joining [28,61,108,221,222,223,224,225,277]. Others have argued

that a postinitiation step is targeted, resulting in either slowed elongation, ribosome

drop-off, or nascent polypeptide degradation [30,201,228,229,230].

One factor contributing to the lack of a consensus model for miRNA function is the

evidence that miRNA targeting of an mRNA significantly reduces message levels

(despite previous reports to the contrary) [26,29,33,61,105,106,188]. Indeed, very

recent studies from Baek et al. and Selbach et al. found that the changes in mRNA

abundance are not only correlated with the repression of many targets, but also can

account for most of the observed reduction in protein expression [90,91]. mRNA

targets of the same miRNA can either be translationally repressed with little change in

mRNA abundance, translationally repressed and have concordant changes in mRNA

abundance, or have little translation repression with large changes in mRNA

abundance [61,101,235]. That miRNAs can affect both protein production and

abundance of their mRNA targets raises the question of to what extent these outcomes

of miRNA regulation are mediated by a common mechanism or by competing or

complementary processes. The regulatory consequence of a particular miRNA–mRNA

interaction might be influenced by miRNA-independent factors such as cellular

context or by additional information encoded by the target mRNA, e.g., presence of

binding sites for other RNA-binding proteins and miRNAs, secondary structure

around miRNA binding sites, or the intrinsic decay rate of the mRNA

[99,107,236,237].

Experiment-specific effects of in vitro translation assays, reporter constructs, or

procedural differences that alter properties of gene expression could account for some

of the wide variation in the apparent mechanisms by which miRNAs alter expression

[99,111]. To date, most studies on translational regulation by miRNAs have used

reporter assays. Although assays that rely on engineered reporter transcripts are

powerful, assay-specific anomalies are a concern; artificial mRNAs may lack key

pieces of regulatory information, overexpression of reporter mRNAs could mask

91

subtle regulatory functions, and DNA transfection can lead to indirect effects on cell

physiology [110]. Indeed, recent reports have found that differences in experimental

setup, such as the method of transfection, type of 5′-cap, or the promoter sequence of

the DNA reporter construct can drastically alter the degree or even the apparent mode

of regulation by miRNAs [201,238]. In addition, some models have been based on

studies in which only one or a few targets were studied, which introduces the

possibility of generalizing the behavior of a single miRNA–mRNA interaction that

may not represent the dominant biological mechanism.

Two recent studies avoided many of these caveats by overexpressing, inhibiting or

deleting specific miRNAs and systematically measuring changes in endogenous

mRNA and protein levels using DNA microarrays and stable isotope labeling with

amino acids in cell culture (SILAC), respectively [90,91]. Both studies found mostly

concordant changes in mRNA levels and protein levels, with changes in mRNA levels

accounting for much, but not all, of the changes in protein abundance. With data for

hundreds of endogenous targets, these studies were the first to provide genome-wide

evidence that mRNA degradation accounts for much of the reduction in protein levels.

And whereas these results suggest that translation inhibition accounts for some of the

observed changes in protein abundance of miRNA targets, they do not provide direct

evidence of this, nor do they provide insight into which steps in translation are

regulated, the extent this regulation contributes to reduced gene expression of specific

mRNAs, or its possible links to mRNA decay.

To investigate how miRNAs regulate gene expression, we systematically identified

direct targets of the miRNA miR-124 by measuring the recruitment of target mRNAs

to Argonaute (Ago) proteins, the core components of the miRNA effector complex, as

previously described [200,239,263]. We then measured, in parallel, mRNA abundance

and two indicators of translation rate, ribosome occupancy and ribosome density, for

more than 8,000 genes, using DNA microarrays and a novel polysome encoding

scheme. This strategy allowed us to directly investigate the behavior of miRNA–

mRNA pairs with respect to both mRNA fate and translation, on a genomic scale.

92

Results

Systematic Identification of mRNAs Recruited to Argonautes by miR-124

To study the effects of miR-124 on expression of mRNA targets, we first had to

identify those targets. Recruitment to Ago complexes in response to the expression of

a particular miRNA appears to be the most reliable criterion for target identification

[239]. To this end, we lysed human embryonic kidney (HEK) 293T cells transfected

with miR-124 and isolated Ago-associated RNA by immunopurification (IP) using a

monoclonal antibody that recognizes all four human Ago paralogs [278]. We

measured mRNA enrichment in Ago IPs by comparative DNA microarray

hybridization of samples prepared from immunupurified RNA and total RNA from

cell extracts. Three replicates of Ago and control IPs were performed from both miR-

124 and mock-transfected cells.

To examine the enrichment profiles of the IPs, we first clustered the microarray results

by their similarity and visualized the results as a heatmap, with the degree of

enrichment of each RNA shown on a green (least enriched) to red (most enriched)

scale (Figure S1). The Ago IP enrichment profiles were reproducible as evidenced by

an average Pearson correlation coefficient between mRNA enrichment profiles of Ago

IPs in mock-transfected cells and miR-124–transfected cells of 0.90 and 0.94,

respectively.

Thousands of mRNAs were reproducibly enriched in the Ago IPs from mock-

transfected cells (Figures S1 and S2, and Text S1). We found that the presence of

sequence matches to two highly expressed microRNA families, miR-17-

5p/20/92/106/591.d and miR-19a/b, in the 3′-untranslated regions (UTRs) of mRNAs

significantly correlated with Ago IP enrichment (Text S2), suggesting that association

with Ago is in large part a reflection of the relative occupancy of each mRNA with the

suite of miRNAs endogenously expressed in HEK293T cells. High-confidence Ago-

associated mRNAs (at least 4-fold enriched over the mean, 1,363 mRNAs)

disproportionately encode regulatory proteins (409, p = 0.001), with roles including

93

“transcription factor activity” (95, p = 0.01), “signal transduction” (230, p = 0.02) and

“gene silencing by RNA” (7, p = 0.02).

To identify RNAs specifically recruited to Agos by miR-124, we compared the mRNA

enrichment profiles of Ago IPs from miR-124–transfected cells to Ago IPs from

mock-transfected cells using the significance analysis of microarrays (SAM) modified

two-sample unpaired t-test. At a stringent 1% local false-discovery rate (FDR)

threshold, we identified 623 distinct mRNAs significantly enriched in Ago IPs from

lysates of miR-124–transfected cells compared to Ago IPs from mock-transfected cells

(Figure 1A).

Previous work established that the 5′-end of the miRNA, the “seed region,” is

particularly important for interactions with mRNA targets [29,60,83,89,196,199,204].

In most cases, there is a 6–8 bp stretch of perfect complementarity between the seed

region of the miRNA and a “seed match” sequence in the 3′-UTR of the mRNA

[29,60,83,89,196,199,204]. We reasoned that if the mRNAs specifically recruited to

Agos by miR-124 transfection were physically associated with miR-124, seed match

sequences would be significantly enriched in miR-124–specific IP targets compared to

nontargets. Indeed, we found strong enrichment of 6–8 base seed matches to miR-124

in the 3′-UTRs of miR-124 Ago IP targets (Figure 1B). We also found enrichment

within the coding sequences of miR-124 Ago IP targets, as previously reported (Figure

1B) [29,90,91,200,239,279,280]. For instance, 60% of miR-124 Ago IP targets contain

a perfect match to positions 2–8 of miR-124 (called 7mer-m8) in their 3′-UTRs,

compared to 10% of nontargets (p < 10−185

, hypergeometric distribution), and 23% of

miR-124 Ago IP targets contain a perfect match to positions 2–8 of miR-124 in their

coding sequence, compared to 10% of nontargets (p < 10−23

). After removing mRNAs

with 7mer seed matches in their 3′-UTRs, the remaining miR-124 IP targets were still

significantly, albeit weakly, enriched for 3′-UTR 6mer matches to miR-124 (6mer 2–

7, p = 0.008, 6mer 3–8, p < 10−5

). These data argue that most miR-124 Ago IP targets

were recruited to Agos by direct association with miR-124, via seed matches in their

3′-UTRs or coding sequences.

94

Figure 1. miR-124 Recruits Hundreds of Specific mRNAs to Argonautes.

(A) Supervised hierarchical clustering of the enrichment profiles of putative miR-124 Ago IP targets

(1% local FDR) in Ago IPs from miR-124–transfected cells (blue) and mock-transfected cells (black).

Rows correspond to 789 sequences (representing 623 genomic loci with a Refseq sequence), and

columns represent individual experiments.

(B) Enrichment of seed matches to miR-124 in the 3′-UTRs and coding sequences of miR-124 Ago IP

targets (1% local FDR). The significance of enrichment of seed matches in Ago IP targets was

measured using the hypergeometric distribution function.

95

Systematic Measurement of mRNA Translation Profiles

The standard approach to assess translation in vivo has been the analysis of “polysome

profiles.” After treatment with cycloheximide to trap elongating ribosomes, mRNAs

with no associated ribosomes and those with varying numbers of ribosomes bound can

be separated by velocity sedimentation through a sucrose gradient. The polysome

profile of a gene‟s mRNA provides information on two key parameters in translation:

(1) the fraction of the mRNA species bound by at least one ribosome, and presumably

undergoing translation, referred to as “ribosome occupancy,” and (2) the average

number of ribosomes bound per 100 bases of coding sequence to mRNAs that have at

least one bound ribosome, referred to as the “ribosome density.”

We previously developed a method to systematically measure ribosome occupancy

and ribosome density by measuring the relative amount of each gene‟s mRNA in each

fraction of a polysome profile using DNA microarray hybridization [281]. We have

since developed and implemented a more streamlined approach that uses one DNA

microarray hybridization to measure ribosome occupancy and only a single additional

microarray hybridization to measure ribosome density (Figure 2 and Figure S3). We

measured ribosome occupancy by first pooling ribosome bound fractions and unbound

fractions and adding exogenous doping control RNAs to each (Figure 2A). Poly(A)

RNA from bound and unbound pools was isolated, amplified, coupled to Cy5 and Cy3

dyes, respectively, and comparatively hybridized to DNA microarrays. The ribosome

occupancy for each gene‟s mRNA was obtained after scaling the microarray data

using the doping controls (see Materials and Methods for details).

We determined the ribosome density for each gene‟s mRNA by a “gradient encoding”

strategy in which a graded ratio of each fraction from the ribosome bound fractions

was split into a “heavy” and a “light” pool, respectively. For instance, 99% of the first

fraction (~one ribosome bound) was added to the light pool and 1% was added to the

heavy pool. Then, 98% of the

96

Figure 2. Systematic Translation Profiling by Microarray Analysis.

(A) Schematic of the procedure used to systematically measure each gene‟s mRNA ribosome

occupancy (see Results and Materials and Methods for details). (B) Schematic of the „„gradient

encoding‟‟ method to measure the average number of ribosomes bound to each gene‟s mRNAs (see

Results and Materials and Methods for details).

97

second fraction (1.5–2 ribosomes bound) was added to the light pool and 2% was

added to the heavy pool, and so on, such that the light pool was enriched for mRNAs

associated with fewer ribosomes, and the heavy pool was enriched for mRNAs

associated with a greater number of ribosomes (Figure 2B). The RNA in each pool

was amplified, labeled with Cy5 or Cy3, mixed, and comparatively hybridized to

DNA microarrays. Thus, the Cy5/Cy3 ratio measured at each element on the array is a

monotonic function of the mass-weighted average sedimentation coefficient of the

corresponding mRNA, which is primarily determined by the number of ribosomes

bound to it. The validity of this approach is supported by the very strong concordance

between ribosome density measurements in yeast obtained with the gradient-encoding

method and our previously published ribosome density measurements obtained using

the traditional approach of analyzing each fraction on separate DNA microarrays

(Pearson r = 0.95) (unpublished data) [281]. Further details of the methodology, as

well as control experiments and additional analyses, will be described elsewhere.

To measure the effects of miR-124 on translation, we performed translation profiling

on cell extracts generated from the same miR-124–transfected, or mock-transfected

cell cultures that were used for Ago IPs and mRNA expression profiling (see below).

We obtained high-quality ribosome occupancy and ribosome density measurements on

16,140 sequences (representing 10,455 genes) from three independent mock-

transfected cultures and two miR-124–transfected cultures. There was a strong

concordance between replicate experiments for both the ribosome occupancy and

ribosome number/density measurements, both in terms of the correlation of the gene-

specific measurements (Pearson correlation for ribosome occupancy = 0.85–0.89,

ribosome number = 0.91–0.97) and the means (mean ribosome occupancy = 0.83–

0.87, mean ribosome number = 5.6–6.1 per mRNA), which were derived

independently for each experiment based on the exogenous doping controls.

The measurements from mock-transfected cells provide some general insights into the

translation regulatory program in proliferating human cells. Here, we focus on 8,385

genes that correspond to a Refseq mRNA for which we obtained high-quality

98

measurements in both Ago IP and mRNA expression DNA microarray experiments.

The average ribosome occupancy for the mRNAs from these 8,385 genes was 85%

(25th and 75th quartiles = 0.81 and 0.94, respectively) (Figure 3A) suggesting, that for

most genes, most polyadenylated mRNAs are associated with ribosomes under these

growth conditions and that there are not abundant pools of polyadenylated mRNAs in

an untranslated “compartment.” For more than 97% of the genes analyzed, a majority

of the transcripts were associated with ribosomes; mRNA transcripts of 3% of these

genes (224) were predominantly unassociated with ribosomes (ribosome occupancy

<50%). The reason for the relative exclusion of this small set of mRNAs from the

highly translated pool remains to be determined: possibilities include sequestration

from the translation machinery or a relatively short half-life that results in these

mRNAs spending a correspondingly small fraction of their lives in the translated pool.

We searched for common biological themes among these non–ribosome-associated

mRNAs using gene ontology (GO) term analysis, and found that an unexpectedly

large fraction of these mRNAs encode proteins involved in “regulation of

transcription” (64, p < 10−7

). On the flip side, there were 342 genes whose mRNAs

were almost completely (98% or greater) associated with ribosomes. Many of these

mRNAs encoded proteins involved in metabolism and gene expression, including

“oxidative phosphorylation” (21, p < 10−10

), “nuclear mRNA splicing” (23, p < 10−5

),

“proteasome complex” (11, p = 0.0003) and “glycolysis” (10, p = 0.0002). mRNAs

with low ribosome occupancy (less than 50%) were significantly less abundant than

mRNAs with high ribosome occupancy (greater than 98%) (Kolmogorov-Smirnov

test, p < 10−15

), consistent with the hypothesis that a lower rate of decay, and hence a

greater fraction of the lifespan spent in the translated pool, contributes to ribosome

occupancy.

The average ribosome density for the 8,385 genes with technically high-quality data

across this set of experiments was 0.53 ribosomes per 100 nucleotides (nts) (25th and

75th quartiles = 0.27 and 0.67, respectively), which corresponds to one ribosome per

189 nts (Figure 3B). Given that ribosomes are believed to span ~30 nts of the mRNA,

99

Figure 3. Analysis of Ribosome Occupancy and Ribosome Density in HEK293T

Cells.

(A) The number of genes as a function of ribosome occupancy. The average ribosome occupancy is

85%.

(B) The number of genes as a function of ribosome density. On average, there is one ribosome per 189

nucleotides.

(C) Ribosome density as a function of a gene‟s coding sequence length. Each gene is indicated by a

blue circle. The red line indicates the moving average density value (window of 50). The inset shows

only genes with coding sequences that are shorter than 1,000 nucleotides.

100

the average ribosome density would be approximately one sixth of the maximal

packing density [282]. This spacing suggests that translation initiation is rate limiting

for most mRNAs.

We previously observed a strong negative correlation between an mRNA‟s ribosome

density and its coding sequence length in yeast cells rapidly growing in rich medium

[281]. Subsequent experiments suggested that this relationship is due to either a strong

inverse correlation between initiation rate and coding sequence length [283], or a

decrease in ribosome density as a function of position along the mRNA [284]. We

found the same inverse relationship between the size of a coding sequence and

ribosome density in proliferating mammalian cells (Spearman r = −0.90) (Figure 3C).

Sucrose gradient sedimentation did not clearly resolve polysomes containing more

than seven ribosomes, so it is possible that our method underestimates the number of

ribosome bound to mRNAs with long coding sequences, which could, in principle,

lead to a spurious negative correlation between coding sequence length and ribosome

density. However, the inverse relationship between coding sequence length and

ribosome density is still readily evident when only mRNAs with coding sequences less

than 1,000 nts are considered (r = −0.73), strongly supporting the validity of this

relationship.

These broad similarities between translational programs in proliferating HEK293 cells

and proliferating S. cerevisiae grown in rich medium, suggest that the overall

organization of the program, and perhaps some of the fundamental mechanisms

underlying the regulation of translation, may be similar in rapidly growing yeast and

human cells [281].

mRNA Recruitment to Argonautes by miR-124 Leads to Modest Decreases in

Abundance and Translation Rate

To measure the effects of miR-124 on mRNA expression levels, we profiled mRNA

expression in the same cell cultures that we used for the Ago IPs and translation

profiling. We obtained high-quality measurements for 15,301 genes from three

101

independent mock-transfected cultures and three independent miR-124–transfected

cultures. There was strong concordance between replicate experiments (Pearson r =

0.95–0.97).

To study the effects of miR-124 on the expression of its mRNA targets, we first

compared the changes in mRNA abundance of Ago IP targets of miR-124 (560

mRNAs; 1% local FDR) and nontargets (7,825 mRNAs) between cells transfected

with miR-124 and cells that were mock-transfected. Samples were taken 12 h after the

respective treatments. We plotted the cumulative distributions of miR-124–dependent

Ago IP targets (Figure 4A, green curve) and nontarget mRNAs (Figure 4A, black

curve) as a function of the differences in their mRNA abundance between miR-124

and mock-transfected cells. miR-124 target mRNAs were much more likely to

decrease in abundance after miR-124 transfection than nontargets (p < 10−173

, one-

sided Kolmogorov-Smirnov test). For example, 74% of miR-124 IP targets decreased

at least 15% at the mRNA level, compared to 13% of nontargets. The average

abundance of miR-124 Ago IP targets decreased by 35% compared to nontargets

(Figure 4C, green bar in left panel). The results are consistent with miRNAs having

significant, but modest effects on the mRNA levels of most of their endogenous

mRNA targets.

Previous work has established that perfect seed matches to the miRNA in 3′-UTRs are

important to elicit effects on mRNA abundance [29,60,83,89,196,199,204]. To test the

importance of 3′-UTR seed matches on the expression of miR-124 targets, we plotted

the cumulative distributions of miR-124 IP targets with at least one 7mer 3′-UTR seed

match (379, Figure 4A, red curve) and miR-124 IP targets that lacked a 7mer 3′-UTR

seed match (181, Figure 4A, blue curve). We found that mRNA targets with 7mer 3′-

UTR seed matches were more likely than targets that lacked a 7mer 3′-UTR seed

match to decrease in abundance in the presence of miR-124 (90% of miR-124 IP

targets with a 3′-UTR seed match decreased at least 15%, compared to 49% of targets

that lacked a 7mer 3′-UTR seed match). On average, IP target mRNAs with 7mer 3′-

UTR seed matches decreased 40%, whereas IP targets that did not have a 7mer seed

102

Figure 4. miR-124 Negatively Regulates the Abundance and Translation of

mRNA Targets.

(A) Cumulative distribution of the change in mRNA levels following transfection with miR-124

compared to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) (560, green), IP

targets with at least one 3′-UTR7mer seed match (379, red), IP targets that lacked a 3′-UTR 7mer seed

match (181, blue), and nontargets (7,825, black). mRNA levels of miR-124 Ago IP targets were more

likely than nontargets to decrease following transfection with miR-124 (p < 10−173

).

(B) Cumulative distribution of the change in the estimated translation rate following transfection with

miR-124 compared to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) (green),

IP targets with at least one 3′-UTR7mer seed match (red), IP targets that lacked a 3′-UTR 7mer seed

match (blue), and nontargets (black). Translation rates of miR-124 Ago2 IP targets were more likely

than nontargets to decrease following transfection with miR-124 (p < 10−61

).

(C) Barplot showing the average change in mRNA abundance (left) and translation rate (right)

following transfection with miR-124 of all miR-124 Ago IP targets (green), IP targets with at least one

3′-UTR 7mer seed match (red), IP targets that lacked a 3′-UTR 7mer seed match (blue). The average

change in mRNA abundance and translation of targets was calculated by subtracting the average change

of nontargets for the mRNA abundance and translation rates after transfection with miR-124.

103

match in their 3′-UTR decreased 17%, compared to nontargets (Figure 4C, left

panel).These results underscore the importance of 3′-UTR seed matches for regulation

at the mRNA level, but also demonstrate that a large fraction of miR-124 IP targets

that lack 7mer seed matches to miR-124 in their 3′-UTR are nevertheless regulated at

the mRNA level by miR-124.

To study the effects of miR-124 on translation of targeted mRNAs, we estimated the

change in the translation rates between miR-124-transfected and mock-transfected

cells (ΔTr) for each mRNA as:

124 124 rr

mock mock r

O D ET

O D E

, (1)

where multiplying O, the fraction of the mRNA that is ribosome-bound (ribosome

occupancy), by D, the average number of ribosomes per 100 nts for bound mRNAs

(ribosome density) provides the weighted ribosome density for each mRNA; Er is an

unmeasured value for the elongation rate of any given mRNA and was assumed not to

change (discussed further below). Values Tr obtained from miR-124 transfected cells

were divided by those from mock-transfected cells to estimate the change. We plotted

the cumulative distribution of Tr for miR-124 Ago IP targets and nontargets (Figure

4B). miR-124 targets (Figure 4B, green curve) were much more likely to decrease in

translation rate than nontarget mRNAs (Figure 4B, black curve) (p < 10−62

, one-sided

Kolmogorov-Smirnov test). The apparent translation rate of 47% of miR-124 Ago IP

targets, but only 10% of nontargets, decreased by at least 10%. In line with what we

observed for changes in mRNA abundance, miR-124 IP targets with at least one 7mer

seed match in their 3′-UTR were more likely to decrease in translation rate than miR-

124 IP targets that lacked a 7mer 3′-UTR seed match (56% percent of miR-124 IP

targets with a 7mer 3′-UTR seed match decreased at least 10% in translation versus

27% of IP targets that lacked a 7mer 3′-UTR seed match). The overall effects on

translation, while highly significant, were very modest; on average, the estimated

translation rates of miR-124 Ago IP targets decreased by 12% relative to nontargets

104

(15% for miR-124 IP targets with a 7mer 3′-UTR seed match and 5% for miR-124 IP

targets without a 7mer 3′-UTR seed match) (Figure 4C, right panel). These results

show that miR-124 has modest effects on the abundance, translation rate, or both for

most its targets.

In some cases, mRNAs that are translationally-repressed are deadenylated and stored,

rather than degraded [220,285,286]. All of our measurements were of mRNAs

amplified based on their poly(A) tails. Therefore, it was possible that the effects on

translation were underestimated and the effects on abundance were overestimated

because a large percentage of targets mRNAs were translationally repressed, and

stored without a poly(A) tail. To test this possibility, we measured the differences in

total RNA levels irrespective of poly(A) tail for each gene between miR-124

transfected and mock-transfected cells. We found that the differences in RNA

abundance between miR-124 transfected and mock-transfected cells as measured with

unamplified total RNA were similar to those measured for amplified poly(A)-selected

mRNA for miR-124 targets (Pearson r = 0.82, slope of least-squares regression fit in

linear space = 0.82) (Figure S4). These data suggest that the apparent decrease in

abundance of miR-124 target mRNAs results primarily from degradation rather than

deadenylation alone.

miR-124 Affects Both the Ribosome Occupancy and Ribosome Density of

Hundreds of Targets

Many steps in protein synthesis have been proposed to be regulated by miRNAs. The

proposed mechanisms include: (i) blocking initiation, e.g., by preventing eiF4F

binding to mRNA caps or joining of the 40S and 60S ribosomal subunits; (ii)

promoting poly(A) tail deadenylation, which can slow initiation by preventing

interactions between the poly(A) tail and 5′-cap, and by increasing the rate of mRNA

decay, which reduces the fraction of the mRNA‟s lifespan spent in the translated pool;

(iii) promoting premature ribosome release during elongation; (iv) slowing translation

elongation; (v) promoting cotranslational proteolysis; and (vi) concerted slowing of

initiation and elongation

105

[28,30,58,60,61,102,201,217,218,221,222,223,225,228,229,230,237,270,277]. The

first four proposed mechanisms make specific predictions about the effects of

miRNAs on the ribosome occupancy and ribosome density of targets. Proposed

mechanisms (i), (ii), and (iii) predict that both occupancy and density will decrease;

mechanism (iv) predicts that ribosome density will increase as a function of the extent

to which elongation is slowed. In contrast, proposed mechanism (v) does not predict

that ribosome occupancy or ribosome density will change, and the effects on ribosome

occupancy and ribosome density in mechanism (vi) depend on the relative effects of

the miRNA on the two steps.

We tested these predictions by comparing ribosome occupancy and density profiles of

mRNAs from miR-124 and from mock-transfected cells. We found that miR-124 Ago

IP targets were much more likely than nontarget mRNAs to exhibit both reduced

ribosome occupancy (Figure 5A) (p < 10−31

, one-sided Kolmogorov-Smirnov test) and

reduced ribosome density (Figure 5B) (p < 10−51

, one-sided Kolmogorov-Smirnov

test) following miR-124 transfection. Thirty-nine percent of miR-124 Ago IP targets

decreased at least 5% in ribosome occupancy, compared to 13% of nontargets; 55% of

miR-124 Ago IP targets decreased at least 5% in ribosome density, compared to 18%

of nontargets. On average, the ribosome occupancy of miR-124 Ago IP targets

decreased by 4%, and their ribosome density decreased by 8% (Figure 5C, green bars).

We hypothesized that mRNAs with fewer associated ribosomes might exhibit larger

changes in ribosome occupancy as a result of the increased likelihood of losing all

ribosomes. In support of this hypothesis, on average, all ten miR-124 target mRNAs

with ribosome occupancy changes greater than 20% had significantly shorter coding

sequences and fewer bound ribosomes than mRNAs that changed less than 20% (p =

0.0003, one-sided Mann-Whitney test) (Figure S5A).

The effects on ribosome occupancy and ribosome density were significantly larger for

miR-124 Ago IP targets that contain at least one 3′-UTR 7mer seed match (45% and

65% decreased at least 5% in ribosome occupancy and ribosome density,

respectively), compared to miR-124 Ago IP targets that lack a 3′-UTR 7mer seed

106

Figure 5. miR-124 Ago IP Targets Decrease in Ribosome Occupancy and

Ribosome Density Due to the Presence of miR-124.

(A) Cumulative distribution of the change in ribosome occupancy following transfection with miR-124,

compared to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) (green), IP targets

with at least one 3′-UTR7mer seed match (red), IP targets that lacked a 3′-UTR 7mer seed match (blue),

and nontargets (black). Changes in ribosome occupancy of miR-124 Ago IP targets were greater than

those for nontargets (p < 10−31

).

(B) Cumulative distribution of the change in ribosome density following transfection with miR-124

compared to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) (green), IP targets

with at least one 3′-UTR7mer seed match (red), IP targets that lacked a 3′-UTR 7mer seed match (blue),

and nontargets (black). Changes in ribosome density of miR-124 Ago IP targets were greater than those

for nontargets (p < 10−51

).

(C) Bar plot of the average change in ribosome occupancy (left) and ribosome density (right) following

transfection with miR-124 of all miR-124 Ago IP targets (green), IP targets with at least one 3′-UTR

7mer seed match (red), IP targets that lacked a 3′-UTR 7mer seed match (blue). The average change in

107

ribosome occupancy and ribosome density of targets was calculated by subtracting the average change

of nontargets for the mRNA abundance and translation rate measurements following transfection with

miR-124. The error bars represent 95% confidence intervals in the mean difference estimated by

bootstrap analysis.

match (26% and 34% decreased at least 5% in ribosome occupancy and ribosome

density, respectively), providing direct evidence for the general importance of 3′-UTR

seed matches for miRNA-mediated translational repression of endogenous mRNAs

[90,91].

The observed effects on ribosome occupancy and density could, in principle, be the

result of multiple independent regulatory mechanisms. For instance, the decrease in

ribosome occupancy and density could be a result of mechanisms (i), (ii), and (iii). If

however, the effects on ribosome occupancy and ribosome density were due to the

same regulatory mechanism, we would expect a large overlap between mRNAs that

show appreciable decreases in ribosome occupancy and ribosome density in the

presence of miR-124. Indeed, 77% of miR-124 IP targets that decreased at least 5% in

ribosome occupancy also decreased at least 5% in ribosome density (30% of all miR-

124 IP targets decreased at least 5% in both ribosome occupancy and ribosome density

compared to 2% of nontargets), which is significantly more than expected by

chance (p < 10−18

, hypergeometric distribution). There was also a modest, but highly

significant, correlation between changes in ribosome occupancy and ribosome density

of miR-124 Ago IP targets (Spearman r = 0.45, p < 10−25

) (Figure S6A), although

many mRNAs appeared to differentially change in either ribosome occupancy or

ribosome density (some miR-124 mRNA targets even appeared to increase

appreciably in ribosome occupancy; Figure S6 and Text S3). These results are

consistent with the effects on ribosome occupancy and ribosome density arising from

the same regulatory mechanism.

If miR-124 induced ribosome drop-off (mechanism (iii)) stochastically along the

coding sequence, the change in ribosome density would be exponentially related to the

length of the coding sequence. To test this, we plotted the change in ribosome density

108

as a function of mRNA length for miR-124 IP targets and found that although they are

correlated (Spearman r = 0.30), it is highly unlikely there is a first-order exponential

relationship between the change in density and the length of the mRNA‟s coding

sequence (p < 10−211

, F-test with the null hypothesis that the observed change in

density fits the predicted change in density from an exponential least-squares fit)

(Figure S5B). Thus, if ribosome drop-off is the predominant mode of miR-124

regulation, it occurs preferentially near the translation start site.

The observation that many miR-124 targets decreased in both ribosome occupancy

and ribosome density after transfection with miR-124 is consistent with regulation of

translation initiation (mechanisms (i) or (ii)) or ribosome drop-off preferentially near

the translation start site (mechanism (iii)) by miR-124 and suggests that slowed

elongation (model (iv)) is not the predominant mode of regulation of translation by

miR-124 under these conditions. Without measurements of the actual effects on

protein synthesis, these results, however, do not rule out the possibility that miR-124

also induces cotranslational proteolysis (v) or coordinately represses translation

initiation and translation elongation (vi), resulting in modest decreases in ribosome

occupancy and ribosome density, but large effects on protein synthesis.

The Effects of miR-124 Transfection on Protein Products of miR-124 Targets

To analyze the overall effect of the observed decreases in mRNA abundance and

translation on protein production, we calculated the estimated change in protein

synthesis as:

124c r

mock

AP T

A

, (2)

where the estimated change in protein synthesis (Pc) can then be derived by

multiplying the change in mRNA abundance by the estimated change in translation

rate (Tr). The change in relative mRNA abundance is calculated by dividing relative

mRNA abundance values from miR-124 transfection experiments by values from the

mock-condition124

mock

A

A

. Although the overall effect on predicted protein production

109

was on average quite modest (~2-fold decrease compared to nontargets), for a small

fraction of miR-124 targets the predicted changes in protein production were fairly

large; 45 of the 560 identified miR-124 targets were predicted to have a decrease of at

least 4-fold in protein production 12 h after miR-124 transfection. A disproportionate

fraction of the most significantly affected mRNAs encoded proteins associated with

membrane compartments (28, p = 0.001), including endoplasmic reticulum (seven)

and plasma membrane (nine); these mRNAs are likely to be translated on the rough

endoplasmic reticulum. A similar observation was reported with different miRNAs in

a recent study [91]. These results suggest that mRNAs that are translated on the rough

endoplasmic reticulum might be particularly susceptible to miRNA-mediated

regulation, possibly while stalled prior to engagement with the endoplasmic reticulum

[287].

To test whether our estimated changes in protein synthesis predict actual changes in

protein abundance, we measured changes in protein abundance of a diverse set of

proteins encoded by mRNAs that are highly enriched in miR-124 Ago IPs by Western

blot analysis based on the availability of reliable antibodies. We chose 14 proteins

encoded by mRNAs that are highly enriched in miR-124 Ago IPs, with predicted

decreases in protein synthesis ranging from no change to 3-fold (Table S1). We

collected cell lysates 60 h (four to five cell divisions) after miR-124 or mock-

transfection to reduce the likelihood of underestimating the change in protein synthesis

for long-lived proteins. Twelve of the 14 antibodies detected bands at the predicted

molecular weight (Figure 6A). We observed a significant correlation between the

estimated changes in protein synthesis (Figure 6B, x-axis) and the measured changes

in protein levels (Figure 6B, y-axis) in response to miR-124 transfection (Spearman r

= 0.54, p = 0.07, slope of least-squares regression fit = 0.54, grey line in Figure 6B),

with one exception. Only RNF128, with a predicted 3.7-fold reduction in protein

synthesis, drastically disagreed with our measured decrease of 1.2-fold reduction. It is

possible that the discordance in RNF128 protein levels is due to posttranslational

autoregulation, which is common among ring finger proteins [288,289,290]. After

excluding RNF128 from analyses, there is a strong concordance between the two

110

measurements (Spearman r = 0.90, p = 0.0001, slope = 0.95; red line, Figure 6B) for

the remaining 11 proteins. The high correlation and the fact that the slope of the best-

fit line excluding RNF128 is close to one, suggests that miR-124–induced changes in

transcript abundance and translation rate can almost completely account for the

changes in abundance of the targeted proteins. Thus, cotranslation proteolysis

(proposal (v)) and coordinate repression of initiation and elongation (proposal (vi)) are

unlikely to play more than a minor role in miR-124 regulation under these conditions.

Concordant Changes in Abundance and Translation of mRNAs Targeted by

miR-124 Suggests That These Two Regulatory Outcomes Are Functionally

Linked

Multiple distinct miRNA regulatory pathways have been proposed, such that

translational repression and mRNA degradation can be regulated independently, and

these two regulatory consequences are differentially affected by specific features of

miRNA–mRNA interaction [60,62,200,217,236]. The relative magnitude of effects on

translation and decay of targeted mRNAs might be influenced by the sequence context

of the miRNA–mRNA interaction and the particular suite of RNA-binding proteins

associated with the mRNA [55,56,101,214]. If the balance between effects on

translation and effects on decay were influenced in a gene-specific way by features of

the mRNA, we would expect that some targets of miR-124 would have relatively large

changes in translation with little change in mRNA abundance or vice versa. If,

however, miRNA–mRNA interactions act through a single dominant regulatory

pathway that affects both translation and decay, we would expect a strong correlation

between the changes in abundance and translation of mRNA targets of miR-124.

We compared the changes in mRNA abundance (Figure 7, x-axis) to apparent changes

in translation rate (Figure 7, y-axis) for miR-124 Ago IP targets following miR-124

transfection. There was strong correlation between these two regulatory effects

(Pearson r = 0.60, see Text S4 and Figure S7 for estimates of significance of the

correlation), and we found no subpopulation of mRNAs whose translation was

appreciably diminished without corresponding changes in mRNA abundance, and few

111

Figure 6. The Effect of miR-124 Transfection on Protein Production of miR-124

Targets

(A) Western blots of 12 proteins encoded by mRNAs highly enriched in miR-124 Ago IPs from mock-

transfected cells (−) and miR-124 transfected cells (+).The bottom bands are loading controls. The

proteins are arranged according to increasing estimated fold-change in protein synthesis from left to

right.

(B) Scatterplot between estimated changes in protein synthesis (x-axis) and observed changes in protein

levels (y-axis) from Western blots. The gray line is a least-squares linear regression fit of all 12

proteins, and the red line is a least-squares fit of 11 proteins, excluding RNF128 (upper left protein,

shown in blue).

112

mRNAs whose abundance changed significantly without a corresponding change in

translation. To test whether the apparent correlation might be driven solely by mRNAs

with the largest measured changes in abundance and translation, we calculated the

average changes in mRNA abundance and translation in moving windows of ten

mRNAs ranked by their change in mRNA abundance. As shown in Figure 7 (red

curve), we found a persistent, nearly monotonic, relationship between changes in

mRNA abundance and translation that closely matches the least-squares fit of all the

data (Pearson r = 0.91). We obtained similar results when we analyzed miR-124 Ago

IP targets with 7mer 3′-UTR seed matches and those that lacked a 7mer 3′-UTR seed

match (Figure S8), although the correlation was stronger for targets with 7mer 3′-UTR

seed matches (r = 0.60 versus 0.42).

The correlation between changes in mRNA abundance and estimated translation rate,

and the absence of a subgroup of mRNAs regulated at the translational level without

corresponding effects on abundance, is consistent with a model in which these two

regulatory programs are functionally linked. Although there was a measurable

decrease in mRNA abundance for almost all miR-124 targets that significantly

decreased in translation, only about half of the targets that decreased in mRNA

abundance registered a measurable reduction in translation. It is possible that some

mRNA targets are degraded without any appreciable effect on translation (e.g., the

mRNAs are degraded while still associated with ribosomes) or that translation of these

mRNAs is indirectly stimulated in response to miR-124, resulting in no apparent effect

on translation at the time we performed translation assays. Alternatively, as the

changes in translation tended to be smaller than the changes in mRNA abundance, we

may have been unable to accurately measure the small effects on translation of many

targets.

113

Figure 7. Concordant Changes in mRNA Abundance and Translation of miR-124

Ago IP Targets.

Scatterplot between changes in mRNA abundance (x-axis) and the estimated translation rate (y-axis) for

miR-124 Ago IP targets following transfection with miR-124 compared to mock. The gray line is a

least-squares linear regression fit of the data, and the red line is a moving average plot (window of 10).

The slope of the least-squares fit of the data is 0.24 (in linear space, 0.36), and the Pearson correlation is

0.60.

114

Changes in Abundance and Translation of miR-124 Ago IP Targets with Seed

Matches in 3′-UTRs, Coding Sequences, and 5′-UTRs

Although most functional microRNA seed matches are located in 3′-UTRs as judged

by mRNA expression data, phylogenetic conservation analysis, Ago IPs, and reporter

studies, some sites in coding sequences and 5′-UTRs can also confer regulation by

miRNAs [83,88,90,91,200,201,202,239,279,280,291]. The 560 high-confidence miR-

124 Ago IP targets for which we obtained high-quality measurements in expression

and translation analyses were strongly enriched for mRNAs that contained miR-124

seed matches in 3′-UTRs and coding sequences (Figure 1B), but they were also

significantly, albeit weakly, enriched, for seed matches in 5′-UTRs (16, p = 0.009).

We compared the effectiveness of 7mer seed matches in the 3′-UTR, coding sequence,

and 5′-UTR, and 6mer seed matches in the 3′-UTR in effecting changes in mRNA

abundance and estimated translation rate. We found that both the abundance and

translation rate of IP targets, regardless of the location of seed matches, decreased

relative to nontarget mRNAs in miR-124 transfected cells compared to mock-

transfected cells (Figure S9). The estimated effects on protein production were

greatest for mRNAs with 7mer seed matches in the 3′-UTR, consistent with previous

studies reporting that 3′-UTR seed matches confer the highest degree of regulation

[29,88,200,239]. Changes in mRNA abundance were significantly greater than

changes in translation for miR-124 Ago IP targets with 3′-UTR and coding sequence

seed matches (Figure S9). IP targets that did not contain any 6mer seed matches were

also significantly more likely to decrease in mRNA abundance than nontargets (Figure

S9), which suggests that many of these mRNAs are specifically recruited to Agos by

miR-124 and regulated by miR-124, even though they do not appear to have canonical

recognition elements.

115

Efficiency of Recruitment to Argonautes by miR-124 Seed Matches Correlates

with Effects on Both mRNA Abundance and Translation

The extent to which each of the thousands of genes expressed in a given mammalian

cell is regulated by the suite of (often hundreds of) miRNAs expressed in that cell is

not known. We reasoned that our Ago immunopurification strategy, by quantitatively

measuring association of mRNAs with microRNA effector complexes, could serve as

a direct readout of the potency of the regulatory effects of miRNAs on each mRNA.

We compared the change in Ago IP enrichment following transfection with miR-124

to the estimated changes in protein production (equation 2) for mRNAs with seed

matches to miR-124 in their 3′-UTR or coding sequence (Figure S10). For mRNAs

with 7mer or 8mer seed matches to miR-124 in their 3′-UTR, there was a strong

negative correlation between the magnitude of their enrichment by the Ago IP and the

estimated changes in production of the protein they encode (3′-UTR 7mer: Pearson r =

−0.72, p < 10−192

; 3′-UTR 8mer: r = −0.72, p < 10−26

) (Figure S10A). For mRNAs

with 7mer or 8mer seed matches to miR-124 in their coding sequences, but no 7mer

seed matches in their 3′-UTRs, there was also a significant, albeit weaker, correlation

(CODING SEQUENCE 7mer: r = -0.39, p < 10−33

; CODING SEQUENCE 8mer: r =

−0.38, p < 10−4

) (Figure S10B). There was also a weak, but still significant,

correlation between IP enrichment and the estimated change in protein production for

mRNAs that lacked 7mer seed matches in their 3′-UTR or coding sequence or that

lacked even 6mer seed matches in their 3′-UTR or coding sequence, respectively (3′-

UTR 6mer: r= −0.40, p < 10−75

; no 3′-UTR or CODING SEQUENCE 6mer: r =

−0.23, p < 10−24

) (Figure S10C). Most of the mRNAs with 7mer or 8mer seed matches

in their 3′-UTR or coding sequence that decreased significantly in protein production

were enriched in the Ago IPs. Thus, Ago IP enrichment following transfection with a

specific miRNA appears to be a good predictor of the corresponding effects on protein

production. Because changes in mRNA abundance and translation following

transfection of a specific miRNA are quantitatively smaller and less specific than their

change in association with Agos, the IP method appears to be a more sensitive assay to

identify the direct regulatory targets of specific miRNAs.

116

Discussion

miRNAs regulate the posttranscriptional fates of most mammalian mRNAs, yet for

endogenous mRNAs, the effects of miRNAs on translation, the steps in translation that

are regulated by miRNAs, and the relationship between regulation of translation and

mRNA decay by miRNAs have not been systematically explored. To address these

effects and relationships, we determined the effect of a human miRNA, miR-124, on

translation and abundance of hundreds of endogenous mRNAs that were recruited to

Argonaute proteins in response to ectopic expression of miR-124 in HEK293T cells.

We developed a simple and economical method to quantitatively measure two key

parameters of translation, ribosome occupancy and average ribosome density, on a

genome-wide scale with single DNA microarray hybridizations for each (Figure 2).

This method allowed us to address the effects of miR-124 on translation of

endogenous mRNAs; it is also more broadly applicable to the study of translational

regulation. In this initial application, we found many parallels between the translation

programs in proliferating human embryonic kidney cells and S. cerevisiae (Figure 3),

suggesting common features of translational programs in eukaryotes [281]. Direct

identification of the mRNAs specifically recruited by miR-124 to Ago proteins, core

components of miRNA-effector complexes, defined functional targets of this miRNA

in this model system, providing a starting point for dissecting miRNA regulation

[200,239,262,263,292]. mRNA expression profiling then allowed us to recognize the

specific effects of miR-124 on the abundance of these targets.

Three major conclusions emerged from our studies: (i) miR-124 reduces translation

and abundance of its mRNA targets over a broad range; changes in mRNA abundance

accounted for ~75% of the estimated effect on protein production; (ii) miR-124

predominantly targets translation at the initiation stage or stimulates ribosome drop-off

preferentially near the translation start site; and (iii) miR-124–mediated regulation of

translation and mRNA decay are correlated, indicating that most mRNAs are not

differentially targeted for translational repression versus mRNA decay.

117

Transfection of miR-124 consistently reduced the translation and abundance of most

of its several hundred high-confidence targets; the resulting decrease in translation

averaged 12% and the decrease in target mRNA abundance averaged 35% (Figure 4).

The observation that there were several mRNAs (CD164, VAMP3, and DNAJC1) that

had about 10-fold reductions in mRNA levels (Figure S7), and the fact that 90% of

control-transfected cells expressed the transfected GFP marker, suggests that more

than 90% cells were transfected with functionally significant quantities of miR-124;

thus the small magnitude of the effects on translation and abundance of most of the

mRNA targets of miR-124 identified by Ago IP was not likely a result of poor

transfection efficiency. The correlation between predicted changes in protein synthesis

and observed changes in protein levels for 11 of 12 proteins following miR-124

transfection (Figure 6), suggests that our assays capture most (or all) of the effects of

miR-124 on protein synthesis.

Although we need to be cautious in generalizing from these model systems, in these

cells under the condition examined, miRNAs appears to modulate production for

hundreds of proteins through joint regulation of target mRNA translation and stability

over a strikingly large dynamic range. While the repressive effects on most targets

were modest (1–3-fold), there were eight targets (DNAJC1, VAMP3, CD164, SYPL1,

MAGT1, HADHB, ATP6V0E1, and SGMS2) that were substantially down-regulated

with decreases in protein synthesis of 10-fold or greater. In addition, 47 targets were

estimated to have greater than 4-fold changes in protein synthesis. Regardless of the

magnitude of regulation, mRNA destabilization accounted for ~75% of the change in

estimated protein synthesis. This range of regulation is in good accord with previous

studies with genetically characterized endogenous miRNAs as well as with studies

introducing exogenous miRNAs introduced into human tissue culture [8,26,30,90,91].

However, our observation that miR-124 had only modest effects on the translation of

hundreds of targets contrasts dramatically with several previous studies in which

miRNAs reduced protein expression by 5–25-fold while only modestly decreasing

mRNA levels (1.1–2-fold), suggesting substantial inhibitory effects on translation

118

[60,62,108,229,238]. The previous studies, however, measured the effect of a specific

miRNA on reporter constructs in which the 3′-UTRs of the encoded mRNAs were not

derived from mammalian mRNAs, but were either short (~250 nts) modified viral

sequences or artificial. In contrast, mammalian mRNA 3′-UTRs tend to be much

longer (on average ~1,000 nts) and include regulatory sites for RNA-binding proteins

and regulatory RNAs that influence mRNA localization, translation, and decay. The

basis for the discrepancy in the results from these two experimental designs remains to

be determined, and the answer is likely to provide useful mechanistic insights. One

possibility is that mRNAs containing exogenous 3′-UTRs might have anomalously

long mRNA half-lives that obscure the normal contribution of mRNA degradation to

the miRNA-directed inhibition of protein expression. The large magnitude of effects

observed in reporter-based assays, compared to what we and others have observed

with endogenous mRNAs, is likely to be partially due to the multiple (four to eight)

engineered miRNA binding sites in the reporter constructs used in those studies

[60,62,108,229,238]. Further, these sites were in close proximity, and adjacent

miRNA binding sites have been reported to act cooperatively [58,88,197]. Indeed, two

studies that measured the effects of specific miRNAs on protein and mRNA levels of

reporters with endogenous mammalian 3′-UTRs found more modest effects on

translation, less than 2-fold on average [29,263]. Moreover, the magnitude of the

effects we observed on translation of the mRNAs targeted by miR-124 were in

agreement with two recent studies that inferred the repressive effect of miRNAs on

translation by measuring miRNA-mediated effects on mRNA and protein abundance

[90,91]. Those reports, based on directly measured changes in protein levels by

quantitative mass spectrometry, concluded that the effects of miRNAs on translation

were small (less than 2-fold for hundreds of target mRNAs).

Although we believe that our experimental design provided a good model of miRNA

regulation as it normally operates in vivo, our results do represent the full range of

possible regulatory consequences of miRNA–mRNA interactions. Our results suggest

that miRNAs have a large dynamic range of effects on endogenous protein expression,

achieved via regulation of both translation and mRNA abundance; this pattern is

119

generally quite consistent with previous results from cells grown in culture and limited

in vivo observations. However, in specific developmental or physiological programs,

or for specific mRNAs, the effects on abundance and translation, as well as the

apparent mode of translation regulation may differ from what we observed in this

study [8,30,228,230,293]. Thus, the effects we observed for miR-124 targets after

ectopically expressing the microRNA in Hek293T cells may not capture the full scope

of regulation by miRNAs in their endogenous context; miR-124 is endogenously

expressed in neuronal cells, and the regulatory effects of miR-124 interactions may be

modulated by the physiological demands of the cell and the specific suite of specific

RNA-binding proteins and regulatory RNAs that also associate with miR-124 target

mRNAs.

miR-124 negatively affected both the ribosome occupancy and ribosome density of

hundreds of its targets (Figure 6). These parallel effects, combined with the close

match between changes in protein synthesis predicted from miRNA-induced effects on

mRNA abundance and translation and changes in protein levels for 11 of 12 proteins,

suggest that the step in translation principally targeted by miR-124 and presumably

other miRNAs is initiation or elongation processivity near the translation start site. We

favor the initiation model because it is in accord with several previous studies that

focused on one or a few mRNAs [28,61,108,221,222,223,225,277], and there is a

paucity of empirical evidence supporting ribosome drop-off, which predicts that

ribosome density of miRNA-regulated mRNAs declines between the 5'- and 3'-ends of

the coding sequence and that there should be an overrepresentation of incompletely

synthesized N-terminal nascent polypeptides [229].

The small apparent magnitude of the effects on translation initiation, combined with

the strong correlation between changes in translation and mRNA abundance, can be

explained by a model in which repression of translational by miR-124 rapidly leads to

mRNA decay. Such a model would explain why observable effects on translation

appear to be smaller than the changes in mRNA abundance: if mRNAs whose

translation is inhibited are quickly destroyed, their diminished translation would not be

120

detected in our translation assay. There is already compelling evidence that

translational repression and mRNA decay are linked

[215,231,232,233,234,294,295,296,297,298,299,300,301,302]. Our observation that

an overwhelming majority of polyadenylated mRNAs are associated with ribosomes

in HEK293T cells may be a manifestation of this relationship (Figure 3A). Thus,

miRNA-mediated inhibition of translation may be linked to a general system for

removal of the mRNA from the translational pool, involving recruitment to P-bodies

and subsequent destruction [56,108,194,195,215,216]. Regulated decoupling of

miRNA-mediated translation repression and mRNA decay would then allow

organisms to tilt the balance of effects in favor of translational repression during

physiological and developmental conditions where mRNA destruction is a

disadvantage [293]. Our results are also consistent with a model in which miRNA-

mediated regulation of translation and mRNA decay are functionally independent, but

are similarly controlled by the same cis-elements. Determining whether the concordant

regulation of translation and mRNA abundance represents a mechanistic coupling of

miRNA-mediated regulation of translation and mRNA decay, and understanding the

molecular links between these two regulatory consequences of miRNA–mRNA

interactions are important goals for future investigation.

121

Materials and Methods

Plasmids and Oligonucleotides

miR-124 siRNA:

sense: 5′-UAA GGC ACG CGG UGA AUG CCA-3′

antisense: 5′-GCA UUC ACC GCG UGC CUU AAU-3′

Cell Culture and Transfection

HEK293T cells were obtained from ATCC (Cat# CRL-11268) and grown in

Dulbecco‟s modified Eagle‟s medium (DMEM) (Invitrogen) with 10% fetal bovine

serum (Invitrogen) and supplemented with 100 U/ml penicillin, 100 mg/ml

streptomycin, and 4 mM glutamineat 37 °C and 5% CO2. Transfections of HEK293T

cells were carried out with calcium phosphate. Cells were plated in 15 cm dishes 12 h

prior to transfection at 2 × 105 cells per ml (25 ml total). We made mock-transfection

mixture (1/10 volume of growth medium) by diluting 152 µl of 2 M CaCl2 into 1.25

ml of nuclease-free H2O and then slowly adding this solution to 1.25 ml of 2× HBS

(50 mM Hepes [pH 7.1], 280 mM NaCl, 1.5 mM Na2HPO4). After 1 min, the mixture

was added to a 15-cm plate at a medium pace. Transfections with miR-124

oligonucleotides were performed analogously with 30 nM of oligonucleotides in 2.5

ml of transfection mixture.

Preparation of Beads for Immunopurifications

Ago-specific 4f9 hybridoma was grown in suspension and adapted to 10% FBS-

enriched DMEM [278]. We purified the antibody by passing supernatant from 1 l of

culture over a 5-ml protein L-agarose column (Pierce Cat# 89929) as per the vendor‟s

instructions. Eluent fractions were pooled and dialyzed into PBS with Slide-A-Lyzer

Dialysis Cassettes (Pierce Cat# 66382). We then biotinylated the purified 4f9 antibody

with No-Weigh NHS-PEG4-Biotin Microtubes (Pierce Cat# 21329). We quantified

biotinylation with EZ Biotin Quantitation Kit (Pierce Cat# 28005). Biotinylated 4f9

antibody was aliquoted and stored at -80 oC until use. For Ago immunopurifications,

we coupled biotinylated 4f9 antibody to DYNAL Dynabeads M-280 Streptavidin

122

magnetic beads (Invitrogen Cat# 112-06D) (50 μg of antibody per ml of beads) as per

vendor‟s instructions and stored the coupled beads at 4 °C for up to 1 wk before use.

Immunoaffinity Purifications

Twelve hours after transfection, we washed each 15-cm plate once with phosphate-

buffered saline (usually two plates were used per IP), then added 1 ml of 4 °C lysis

buffer (150 mM KCl, 25 mM Tris-HCl [pH 7.4], 5 mM Na-EDTA [pH 8.0], 0.5%

Nonidet P-40, 0.5 mM DTT, 10 μl protease inhibitor cocktail [Pierce Cat# 78437], 100

U/ml SUPERase•In [Ambion Cat# AM2694]). Following a 30-min incubation at 4 °C,

we scraped the plates, combined the lysates, and then spun them at 4 °C for 30 min at

14,000 RPM in a microcentrifuge. We collected the supernatant and filtered it through

a 0.45-µm syringe filter. We froze an aliquot of lysate in liquid nitrogen for reference

RNA isolation. We then added 0.22 mg/ml heparin to the lysate. We mixed the lysate

with 2.5 mg of Dynal m-280 streptavidin beads (250 ul from original storage solution)

coupled to biotinylated 4F9 ago antibody (~12.5 µg), which we equilibrated

immediately prior to use by washing twice with 1 ml of lysis buffer. We incubated the

beads with the lysate for 2 h at 4 °C and then washed the beads twice with 1.25 ml of

ice-cold lysis buffer for 5-min each. Five percent of the beads were frozen for SDS

PAGE analysis after the second wash. RNA was extracted directly from the remaining

beads using lysis buffer from Invitrogen‟s Micro-to-Midi kit (Invitrogen Cat# 12183-

018). We purified RNA from the lysate and RNA extracted from the beads with the

Micro-to-Midi kit as per vender‟s instructions, except that the percentage isopropanol

used for binding to the column was 70%, instead of 33%, to promote the binding of

small RNAs.

Western Blots

Sixty hours after transfection, Hek293T cell lysate was prepared using the same

protocol for immunoaffinity purifications. The concentration of protein in each sample

was quantified using the BCA assay (Pierce Cat#23255). For SDS-PAGE separation,

25 µg of protein from each sample was used. Protein was then transferred on to a

123

polyvinylidene fluoride (PVDF) membrane for detection with the following specific

antibodies:DUSP9 (Abcam Cat# ab54941-100); PTPN11 (Bethyl Laboraties Cat#

a301–544a); ITGB1 (BD Transduction Laboraties Cat# 610467); AURKA, DHCR24,

MAPK14, PLK1 (Cell Signaling Cat# 4718, 2033s, 9212, 4513, respectively); AHR,

ACTN4, CDK4, RNF128, NRAS, PTBP2, (Santa Cruz Biotechnology Cat# sc-5579,

sc-17829, sc-260, sc-101967, sc-519, sc-101183, respectively); TUBA1A (Sigma Cat#

096K4777). GAPDH and TUBB1 (Abcam Cat# ab9484, ab6046) were used as loading

controls to check for lane-specific differences from loading, transfer, and detection

errors. Protein bands were quantified using the BioRad Quantity One software

package.

Preparation of Cell Extracts for Translation Profiling

For translation experiments, two 15-cm dishes of cells (per condition) were seeded,

grown, and transfected as described above. Twelve hours after transfection, high-

purity cyclohexamide (Calbiochem Cat# 239764) was added at a final concentration of

0.1 mg/ml directly into growth media, and the plate was agitated for 1 min at room

temperature. Plates were then placed on ice and washed twice with 10 ml of ice-cold

buffer A (20 mM Tris [pH 8.0], 140 mM KCl, 5 mM MgCl2, 0.1 mg/ml

cycloheximide). After the second wash was aspirated, the plates were tilted and left for

1 min on ice to facilitate removal of excess liquid. Each plate was then washed 1×

with 2 ml of ice-cold buffer A that contained 0.22 mg/ml of heparin. After removal of

excess liquid, cells were scraped from each dish and collected in a 1.5-ml

microcentrifuge tube on ice. Each plate typically yielded about 300 μl (for 600 μl

total) of cells and residual buffer. This mixture was then brought to 1× protease

inhibitor cocktail (Pierce Cat# 78437), 100 U/ml SUPERASin, and 0.5 mM DTT. To

lyse the cells, the cell-buffer mixture was brought to 0.1% Brij 58 (Sigma Aldrich

Cat# P5884-100G) and 0.1% sodium deoxycholate (Sigma Aldrich Cat# D6750-

100G) and vortexed for 1 min. The lysate was subsequently spun at 3,500 rpm in a

microcentrifuge for 5 min at 4 °C. Supernatant was collected in a fresh tube and spun

at 9,500 rpm in a microcentrifuge for 5 min at 4 °C. Supernatant was collected, flash

frozen in liquid nitrogen, and then stored at −80 °C until use.

124

Sucrose Gradient Preparation

Sucrose gradients were prepared using the Gradient Master (Biocomp) according to

the manufacturer‟s suggestions. Five percent and 60% (w/v) sucrose solutions were

prepared by dissolving sucrose in Gradient Buffer (20 mM Tris-HCl [pH 8.0], 140

mM KCl, 5 mM MgCl2, 0.5 mM DTT, 0.1 mg/ml cycloheximide) at room

temperature. The 60% solution was dispensed into an SW41 ultracentrifuge tube

through a cannula underneath the 5% solution. Using an 11-step program (Biocomp,

SW41 SHORT SUCR 5-50 11), the two solutions were mixed on the Gradient Master

to form a linear gradient. After preparation, gradients were placed in chilled SW41

ultracentrifuge buckets and equilibrated for several hours at 4 ºC.

Sucrose Gradient Velocity Sedimentation

Immediately before centrifugation, 300 μl of lysate (~300 μg of total RNA) was

transferred to the surface of the gradient. Gradients were centrifuged at 41,000 rpm

(RCFave = 207,000) for 70 min at 4 ºC using a SW41 rotor and then stored at 4 ºC

until fractionation. The Gradient Station (Biocomp) trumpet tip was pushed into the

ultracentrifuge tube at a rate of 0.17 mm per second. Fractions (550 μl) were collected

into a 96-well plate containing 600 μl of lysis solution (Invitrogen) using a fraction

collector (Teledyne-Isco). The absorbance of the gradient at 260 nm was measured

during fractionation using a UV6 system (Teledyne-Isco).

Gradient Encoding

Immediately after fractionation a unique set of four to five polyadenylate-tailed

control RNAs, corresponding to Methanococcus jannaschii mRNAs that do not share

significant identify to sequences in the human genome, were added at 100 pg each to

fractions that contained the 80S ribosome and polysomes. The solution was mixed

well by inverting the plate several times, and liquid was collected in the well bottom

by a brief centrifugation. A Precision XS liquid handler (BioTek Intruments) was used

to transfer a defined volume of each of the fractions to one of four tubes (Fisher Cat#

14-959-11B); the solutions in each tube are referred to as pool, “A,” “B,” “C,” or “D,”

125

respectively. Upon completion of liquid handling, eight additional control RNAs

(Ambion Cat# 1780) were added to each pool, and the pools were stored at −20 ºC.

Pools A–D were thawed at room temperature for 30 min. Two volumes of isopropanol

was added to each pool, and the RNA in each pool was isolated from the mixture

using the Micro-to-Midi RNA isolation kit (Invitrogen Cat# 12183-018).

DNA Microarray Production and Prehybridization Processing

HEEBO oligonucleotide microarrays were printed on epoxysilane-coated glass (Schott

Nexterion E) by the Stanford Functional Genomic Facility. The HEEBO microarrays

contain ~45,000 70-mer oligonucleotide probes, representing ~30,000 unique genes. A

detailed description of this probe set can be found at

(http://microarray.org/sfgf/heebo.do) [255,303].

Prior to hybridization, slides were first incubated in a humidity chamber (Sigma Cat#

H6644) containing 0.5× SSC (1× SSC = 150 mM NaCl, 15 mM sodium citrate [pH

7.0]) for 30 min at room temperature. Slides were snap-dried at 70–80 oC on an

inverted heat block. The free epoxysilane groups were blocked by incubation with 1M

Tris-HCl (pH 9.0), 100 mM ethanolamine, and 0.1% SDS for 20 min at 50 °C. Slides

were washed twice for 1 min each with 400 ml of water, and then dried by

centrifugation. Slides were used the same day.

DNA Microarray Sample Preparation, Hybridization, and Washing

Amplified RNA was used for most DNA microarray experiments. Poly-adenylated

RNAs were amplified in the presence of aminoallyl-UTP with Amino Allyl

MessageAmp II aRNA kit (Ambion Cat# 1753). For mRNA expression experiments,

universal reference RNA was used as an internal standard to enable reliable

comparison of relative transcript levels in multiple samples (Stratagene Cat# 740000).

Amplified RNA (3–10 μg) was fluorescently labeled with NHS-monoester Cy5 or Cy3

(GE HealthSciences Cat# RPN5661). Dye-labeled RNA was fragmented (Ambion

Cat# 8740), then diluted into in a 50-μl solution containing 3× SSC, 25 mM Hepes-

NaOH (pH 7.0), 20 μg of human Cot-1 DNA (Invitrogen Cat# 15279011), 20 μg of

126

poly(A) RNA (Sigma Cat# P9403), 25 μg of yeast tRNA (Invitrogen Cat# 15401029),

and 0.3% SDS. The sample was incubated at 70 °C for 5 min, spun at 14,000 rpm for

10 min in a microcentrifuge, then hybridized at 65 °C using the MAUI hybridization

system (BioMicro) for 12–16 h. For translation experiments, amplified RNA from

pools A and C was fluorescently labeled with NHS-monoester Cy5, and RNA from

pools B and D was fluorescently labeled with NHS-monoester Cy3. Amplified RNA

from pools A and B were comparatively hybridized to a DNA microarray to obtain the

average ribosome density, and amplified RNA from pools C and D were

comparatively hybridized to a DNA microarray to measure ribosome occupancy.

To compare total RNA levels in miR-124 and mock-transfected cells (Figure S3), 5–

10 μg of total RNA from miR-124–transfected cells or mock-transfected cells or

universal reference RNA (Stratagene Cat# 740000) was reverse transcribed with

Superscript III (Invitrogen Cat# 18080085) in the presence of aminoallul-dUTP 5-(3-

aminoallyl)-dUTP (Ambion Cat# AM8439) and natural dNTPs (GE Healthsciences

Cat# US77212) with 10 μg of N9 primer (Invitrogen). Subsequently, amino-allyl–

containing cDNAs from miR-124 and mock-transfected cells were covalently linked

to Cy5 NHS-monoesters, and universal reference cDNA was covalently linked to Cy3

NHS-monoesters (GE HealthSciences Cat# RPN5661). Cy5- and Cy3-labeled cDNAs

were mixed and diluted into 50 μl of solution containing 3× SSC, 25 mM Hepes-

NaOH (pH 7.0), 20 μg human Cot-1 DNA (Invitrogen Cat# 15279011), 20 μg of

poly(A) RNA (Sigma Cat# P9403), 25 μg of yeast tRNA (Invitrogen Cat# 15401029),

and 0.3% SDS. The sample was incubated at 95 °C for 2 min, spun at 14,000 rpm for

10 mins in a microcentrifuge, then hybridized at 65 °C for 12–16 h.

Following hybridization, microarrays were washed in a series of four solutions

containing 400 ml of 2× SSC with 0.05% SDS, 20058 SSC, 1× SSC, and 0.2× SSC,

respectively. The first wash was performed for 5 min at 65 °C. The subsequent washes

were performed at room temperature for 2 min each. Following the last wash, the

microarrays were dried by centrifugation in a low-ozone environment (<5 ppb) to

prevent destruction of Cy dyes [265]. Once dry, the microarrays were kept in a low-

127

ozone environment during storage and scanning (see

http://cmgm.stanford.edu/pbrown/protocols/index.html).

Scanning and Data Processing

Microarrays were scanned using AxonScanner 400oB (Molecular Devices). PMT

levels were autoadjusted to achieve 0.1–0.25% pixel saturation. Each element was

located and analyzed using SpotReader (Niles Scientific) and GenePix Pro 6.0

(Molecular Devices). For IP and mRNA expression experiments, the microarrays were

submitted to the Stanford Microarray Database for further analysis [266]. Data were

filtered to exclude elements that did not have one of the following: a regression

correlation of ≥0.7 between Cy5 and Cy3 signal over the pixels compromising the

array element, or an intensity/background ratio of ≥3 in at least one channel.

Ribosome density (pool “A” versus “B”) and ribosome occupancy (pool “C” versus

“D”) measurements were normalized using exogenous doping control RNAs to correct

for experimental variation between the two pools from RNA isolation, labeling

efficiency, and scanning levels. In most cases, oligonucleotides that were designed to

measure the exogenous doping control RNAs were represented multiple times on the

DNA microarray (up to eight) and printed from different plates with different print

tips. For ribosome occupancy experiments, the measured Cy5/Cy3 ratios of features

on the microarray that correspond to the eight RNA controls added to pools “C” and

“D” were fit to their expected Cy5/Cy3 ratios using least-squares linear regression in

the statistical computing program R. The slope and y-intercept were used to rescale the

measured Cy5 value of all other features on the DNA microarray. The ribosome

occupancy for each RNA was then calculated as the corrected Cy5 intensity /

(corrected Cy5 intensity + Cy3 intensity) (Figure S3C).

To calculate the average number of ribosomes bound to each mRNA, the measured

Cy5/Cy3 ratios of features on the microarray that correspond to the 85 M. jannaschii

doping control RNAs that were added to fractions that contained ribosomes pools was

fit to their expected Cy5/Cy3 ratios using least-squares linear regression. The slope

128

and y-intercept were used to rescale the measured Cy5 value of all other features on

the DNA microarray (Figure S3B).

The average ribosome density was calculated by dividing the average ribosome

number by coding sequence length and then multiplying the result by 100 to give

density per 100 nts. The average ribosome number was calculated using two

relationships. For each ribosome peak in the profile, the distance traveled from the

start point was determined. In all gradients, we could clearly resolve peaks for up to

seven bound ribosomes, and we used least-squares regression to relate the ribosome

peaks 1–7 to their distance traveled in the gradient according to the following

equation:

bDTaRLog )(10 , (3)

where R represents the number of ribosomes bound, DT represents the distance

traveled, and a and b are the slope and y-intercept, respectively. We then recorded the

distance between the midpoint of each fraction to the start of the profile for each of the

15 ribosome-bound fractions and used the slope and y-intercept from Equation 3 to

calculate the number of ribosomes at each fraction midpoint. The gradient encoding

ratio at each fraction midpoint is the result of differential partitioning of each fraction

in a predetermined manner into the heavy and light pools, and the ratio can be related

to the ribosome number at each fraction midpoint using least-squares linear regression

as described by Equation 4:

10 10( ) ( )Log R a Log ER b , (4)

where R represents the average ribosome number, and ER represents the encoding

ratio. Finally, the average number of ribosomes bound for each gene‟s mRNAs was

calculated using the slope and y-intercept from Equation 4.

Prior to normalization, spots with intensity/background of less than 1.5 for either Cy3

or Cy5 channel were filtered.

The microarray data are available from Gene Expression Omnibus (GEO)

(http://www.ncbi.nlm.nih.gov/geo/) and Stanford Microarray Database .

129

Microarray Analyses

Hierarchical clustering was performed with Cluster 3.0 [267] and visualized with Java

TreeView 1.0.12 [268].

For SAM, unpaired two-class t-tests were performed with default settings (R-package

samr; http://cran.r-project.org/web/packages/samr/index.html). Microarray features

that passed quality filtering in all experiments were used as input. Ago IP experiments

and mRNA expression experiments were mean centered at log2 0 prior to running

SAM. The ribosome occupancy and ribosome number/density measurements from

miR-124 and mock-transfected cells were highly correlated, but had slightly different

means (see main text). Because of the small changes in ribosome occupancy and

ribosome density between miR-124 transfected and mock-transfected samples, we

conservatively adjusted the means of each experiment to be the same by subtracting

the difference between the mean of that experiment and the mean of all the

experiments to ensure that differences observed between miR-124–transfected and

mock-transfected cells were not due to the doping control normalization. Enrichment

of GO terms was performed with Genetrail [304]. p-Values were corrected for

multiple-hypothesis testing by the Bonferroni method [305].

The significance of correlations was estimated in R by recalculating the correlations

with 10,000 permuted sets of data, then estimating the p-value with the normal

distribution function using the mean and standard deviation from the permuted data.

We used a bootstrap method to estimate 95% confidence intervals for the average

changes in mRNA abundance, estimated translation rate, ribosome occupancy, and

ribosome density (Figures 4 and 5, and Figure S8) of IP targets compared to

nontargets. To do this, we sampled with replacement measurements for each gene

from the mock and miR-124 replicates, respectively, 10,000 times, then calculated the

respective changes between miR-124 IP targets and nontargets for the 10,000

bootstrapped samples.

130

Sequence Data

For molecular features that mapped to genomic loci with an entrezID, the RefSeq

sequence with the longest 3′-UTR was used. In cases with multiple RefSeqs with the

same 3′-UTR length, the one that was alphanumerically first was used. RefSeq 3′-

UTR, coding, and 5′-UTR sequences were retrieved from UCSC genome browser

(hg18) http://genome.ucsc.edu/. Seed match sites in these sequences were identified

with Perl scripts. miR-124 seed matches: 6mer_n2-7 “UGCCUU,” 6mer_n3-8

“GUGCCU,” 7mer-m8 “GUGCCUU,” 7mer-A1 “UGCCUUA,” 8mer

“GUGCCUUA.” In many instances, there were multiple probes on the DNA

microarrays that mapped to the same Refseq. In these cases ,we used the probe that

was most enriched in Ago IPs from miR-124–transfected cells compared to mock-

transfected cells.

Acknowledgments

Professor Edward Chan very kindly provided 4F9 Ago hybridoma. We thank Dan

Klass for advice on immunopurifications, Dr. Julia Salzman for advice on statistics,

Greg Hogan, Dr. Delquin Gong, Ari Firestone, Graham Anderson, Dustin Hite, and

Dr. Jamie Geier Bates for comments on the manuscript, as well as members of the

Brown, Ferrell and Herschlag labs for advice and discussions.

131

Supplementary Figures

Figure S1. mRNA Enrichment Profiles in Ago IPs.

Unsupervised hierarchical clustering of the enrichment profiles in Ago IPs from miR-124–transfected

cells (black), mock-transfected cells (blue), and negative control IPs (no Ago antibody) from both types

of cells (orange). Rows correspond to 16,095 sequences (representing 9,729 genomic loci with a Refseq

sequence), and columns represent individual experiments. Unsupervised hierarchical cluster analysis

segregated the Ago IPs from the negative-control IPs. The Ago IPs were further bifurcated into two

subgroups: miR-124 transfected and mock transfected. There was a correlation between Ago and mock

IPs (r = 0.6), even though ~10-fold less RNA was obtained from mock IPs compared to Ago IPs, and

no protein bands were detectable by protein staining. We speculate that Ago complexes bind the beads

nonspecifically and contribute to the weak background binding (Text S1 & Figure S2).

132

Figure S2. Streptavidin-coated Dynal Beads Weakly Enrich miR-124 Targets

After miR-124 Transfection.

(A) Supervised hierarchical clustering of the enrichment profiles of the 500 most enriched mRNAs in

negative-control IPs from miR-124–transfected cells (blue) compared to mock-transfected cells (black).

Rows correspond to mRNAs, and columns represent individual experiments.

(B) Enrichment of seed matches to miR-124 in the 3′-UTRs of mRNAs nonspecifically associated with

magnetic beads. The significance of enrichment of seed matches in Ago IP targets was measured with

the hypergeometric distribution.

133

Figure S3. Polysome Profiles and Doping Control Fits.

(A) Polysome profile traces from lysates prepared from mock-and miR-124 transfected Hek293T cells.

The vertical gray line indicates the division between unbound and ribosome bound fractions.

(B) Scatterplot between the observed “gradient-encoding” ratios of 33 exogenous RNAs doped into

134

each of the ribosome bound fractions and their predicted ratios. The blue circles show the raw data and

the blue line is the least-squares fit of the raw data, while the red circles and red line represent the

corrected data and fit, respectively.

(C) Scatterplot between the observed ratios of eight exogenous RNAs doped into each ribosome bound

and ribosome unbound fractions and their predicted ratios. The blue circles show the raw data, and the

blue line represents the least-squares fit of the raw data, while the red circles and red line represent the

corrected data and fit, respectively.

135

Figure S4. miR-124 Ago IP Targets Are Likely Destroyed, Rather than

Deadenylated and Stored.

Scatterplot between changes in mRNA abundance for miR-124 Ago IP targets following transfection

with miR-124 compared to mock measured with poly(A) amplified mRNA (x-axis) and cDNA

synthesized from randomly primed total RNA (y-axis). The red line has a slope of one and goes through

the y-axis at zero. The black line is a least-squares fit of the data (slope = 0.82 in linear space, Pearson

correlation log2 [mRNA] = 0.82). This analysis compares 208 miR-124 targets for which we obtained

quality measurements in both experiments.

136

Figure S5. Relationship Between the Coding Sequence Length and Changes in

Ribosome Occupancy and Ribosome Density of miR-124 Ago IP Targets

Following Transfection of miR-124.

(A) Scatterplot between coding sequence length (x-axis) and fold-change in ribosome occupancy (y-

axis) for miR-124 Ago IP targets following transfection with miR-124 compared to mock. The

horizontal gray line denotes a 20% reduction in ribosome occupancy.

(B) Scatterplot between coding sequence length (x-axis) and fold-change in ribosome density (y-axis)

for miR-124 Ago IP targets following transfection with miR-124 compared to mock. The red curve is a

nonlinear least-squares fit of the change in ribosome density following a first-order decay as a function

of coding sequence length.

137

Figure S6. Relationship Between Ribosome Occupancy in Mock-transfected Cells

and Change in Ribosome Occupancy Following Transfection of miR-124.

(A) Scatterplot between changes in ribosome occupancy (x-axis) and ribosome density (y-axis) for miR-

124 Ago IP targets following transfection with miR-124 compared to mock. The gray line is a least-

138

squares linear regression fit of the data (Spearman rank correlation = 0.45), and the red line is a moving

average plot (window of 10).

(B) The logarithm of the ratio of the average ribosome occupancy in miR-124–transfected cells to that

in mock-transfected cells as a function of the average ribosome occupancy in mock-transfected cells

(Spearman rank correlation = −0.78). Black circles correspond to mRNAs that were not enriched by the

Ago IP following miR-124 transfection. Red circles correspond to mRNAs that were enriched by the

Ago IP following miR-124 transfection (1% local FDR). The green curve represents a Lowess

smoothed fit of the data. The gray curve shows the maximum possible increase in ribosome occupancy

in miR-124 cells compared to mock cells.

(C) The ratio in the average ribosome occupancy in miR-124–transfected cells versus mock-transfected

cells minus the Lowess fit of the data (green points in [B]) as a function of the average ribosome

occupancy in mock-transfected cells.

(D) Scatterplot of changes in mRNA abundance (x-axis) versus changes in translation rate (y-axis) for

Ago IP targets following transfection with miR-124. The slope of the least-squares fit of the data is 0.24

(in linear space, 0.36), and the Pearson correlation is 0.60. This is Figure 7 replotted to allow side-by-

side comparison with (E).

(E) Same as in (D), except that the changes in translation rate were obtained using the smoothed-fit

adjusted ribosome occupancy measurements (C). The slope of the least-squares fit of the data is 0.20 (in

linear space, 0.29), and the Pearson correlation is 0.59.

139

Figure S7. Significance of the Correlation Between Changes in mRNA

Abundance and Translation of miR-124 Ago IP Targets.

(A) Histogram of correlations between changes in mRNA abundance and translation for 100,000

permuted sets of miR-124 Ago IP targets. The red curve shows a normal distribution with mean (0.0)

and standard deviation (.04) from the 100,000 permuted sets. The red arrow shows the Pearson

correlation of the actual data (0.60, p < 10−45

).

(B) Moving average plot of the Pearson correlation (window of 500) between changes in mRNA

abundance and translation as a function of enrichment in Ago IPs in miR-124–transfected cells versus

mock-transfected cells (SAM D-score). The horizontal grey line shows the average Pearson correlation

between changes in mRNA abundance and translation across the moving windows (r = 0.09).

(C) Histogram of correlations between changes in mRNA abundance and translation for 10,000

permuted sets of mRNAs that are not miR-124 targets, but have similar distribution in their changes in

mRNA abundance (t-test, p > 0.001) to miR-124 IP targets that change less than 40% in mRNA

abundance. The red curve shows a normal distribution with mean (0.14) and standard deviation (0.04)

from the 10,000 permuted sets. The red arrow shows the Pearson correlation of miR-124 IP targets that

change less than 40% in mRNA abundance (r = 0.30, p < 10−5

).

140

Figure S8. Concordant Changes in mRNA Abundance and Translation of miR-

124 Ago IP Targets with 7mer 3′-UTR Seed Matches and miR-124 Ago IP

Targets that Lack a 7mer 3′-UTR Seed Match.

(A) Scatterplot between changes in mRNA abundance (x-axis) and the estimated translation rate (y-

axis) for miR-124 Ago IP targets with 7mer 3′-UTR seed matches following transfection with miR-124

compared to mock. The gray line is a least-squares linear regression fit of the data, and the black line is

a moving average plot (window of 10). The slope of the least-squares fit of the data = 0.23 (in linear

space = 0.37) and the Pearson correlation = 0.59.

(B) Scatterplot between changes in mRNA abundance (x-axis) and the estimated translation rate (y-axis)

for miR-124 Ago IP targets that lack 7mer 3′-UTR seed matches following transfection with miR-124

compared to mock. The gray line is a least-squares linear regression fit of the data, and the red line is a

moving average plot (window of 10). The slope of the least-squares fit of the data = 0.21 (in linear

space = 0.24) and the Pearson correlation = 0.42.

141

Figure S9. Changes in Abundance and Translation of miR-124 Ago IP Targets

With Seed Matches in 3′-UTRs, Coding Sequences and 5′-UTRs.

(A) Cumulative distribution of the change in mRNA levels following transfection with miR-124

compared to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) with at least one

3′-UTR 7mer seed match, but no coding sequence or 5′-UTR 7mer seed matches (red, 244), IP targets

with at least one 3′-UTR 6mer seed match (green, 47), but no 3′-UTR, coding sequence, or 5′-UTR

7mer seed matches, IP targets with at least one coding sequence 7mer seed match, but no 3′-UTR or 5′-

UTR 7mer seed matches (blue, 70), IP targets that lacked a 6mer seed match in the 3′-UTR, coding

sequence, or 5′-UTR (orange,23), and nontargets (7385, black). This analysis compares Ago IP targets

(red) versus nontargets (black).

(B) Cumulative distribution of the change in translation following transfection with miR-124 compared

to mock. This analysis compares miR-124 Ago IP targets (1% local FDR) with at least one 3′-UTR

7mer seed match, but no coding sequence or 5′-UTR 7mer seed matches (red), IP targets with at least

one 3′-UTR 6mer seed match (green), but no 7mer seed matches in the 3′-UTR, coding sequence, or 5′-

142

UTR, IP targets with at least one coding sequence 7mer seed match, but no 7mer seed match in the 3′-

UTR or 5′-UTR (blue), IP targets that lacked a 6mer seed match in the 3′-UTR, coding sequence, or 5′-

UTR (orange), and nontargets (black). This analysis compares Ago IP targets (red) versus nontargets

(black).

(C) Bar plot of the average change in mRNA abundance (blue) and translation rate (red) of miR-124

Ago IP targets following transfection with miR-124. The average change in mRNA abundance and

translation of targets was calculated by subtracting the average change of nontargets for the mRNA

abundance and translation rate measurements following transfection with miR-124. This analysis

compares miR-124 Ago IP targets (1% local FDR) with at least one 3′-UTR 7mer seed match, but no

coding sequence or 5′-UTR 7mer seed matches, IP targets with at least one 3′-UTR 6mer seed match,

but no 7mer seed matches in the 3′-UTR, coding sequence, or 5′-UTR, IP targets with at least one

coding sequence 7mer seed match, but no 7mer seed match in the 3′-UTR or 5′-UTR, IP targets with at

least one 7mer seed match in the 5′-UTR, but no 7mer seed match in the 3′-UTR or coding sequence,

and IP targets that lacked a 6mer seed match in the 3′-UTR, coding sequence, or 5′-UTR. This analysis

compares Ago IP targets (red) versus nontargets (black). The error bars represent 95% confidence

intervals in the mean difference estimated by bootstrap analysis.

143

Figure S10. Efficiency of Recruitment to Argonautes by miR-124 Seed Matches

Correlates with Effects on Both mRNA Abundance and Translation.

(A) Scatterplot between changes in Ago IP enrichment (x-axis) following transfection with miR-124

compared to mock and estimated changes in protein production (Equation 2) (y-axis) for mRNAs with

either 8mer seed matches (red dots) or 7mer seed matches (blue dots) to miR-124 in their 3′-UTRs. For

8mer seed matches, the slope of the least-squares fit of the log2 data is −0.46 (in linear space, −0.03),

and the log2 Pearson correlation is −0.72. For 7mer seed matches, the slope of the least-squares fit of the

log2 data is −0.39 (in linear space, −0.05), and the log2 Pearson correlation is −0.72.

(B) Same as (A) except for mRNAs with seed matches to miR-124 in their coding sequences and no

7mer seed matches in their 3′-UTRs. For 8mer seed matches, the slope of the least-squares fit of the log2

data is −0.13 (in linear space, −0.008), and the log2 Pearson correlation is −0.39. For 7mer seed

matches, the slope of the least-squares fit of the log2 data = −0.14 (in linear space, −0.02), and the log2

144

Pearson correlation is 0.38.

(C) Same as in (A) except for mRNAs with 6mer seed matches to miR-124 in their 3′-UTR, but no

7mer seed match in their 3′-UTR or coding sequence (red), and mRNAs that lack 6mer seed matches in

their 3′-UTR or coding sequence (blue). For 6mer seed matches, the slope of the least-squares fit of the

log2 data is −0.21 (in linear space, −0.09), and the log2 Pearson correlation is −0.40. For mRNAs

without 6mer seed matches, the slope of the least-squares fit of the log2 data is −0.13 (in linear space,

−0.07), and the log2 Pearson correlation is −0.22.

Table S1. Summary of miR-124 Targets for Western Blot Analysis.

145

Concluding Remarks

Dicer Domain Function

Regardless of the exact role of DUF283, it is satisfying to observe the additive effects

of Dicer domains on proper dicing. The smallest active catalytic unit, RNab, cleaves

dsRNA in an unmeasured manner with no discernable banding pattern, highlighting

the importance of the PAZ domain in Dicer substrate recognition. Addition of the PAZ

to the RNase domains modulates more precise dicing activity; however the products

are of an insufficient size. It is only in the presence of DUF283 and PAZ that the

RNab domains can effectively generate properly sized siRNAs from dsRNA. The

evolution of such meticulous control over such seemingly minor details belies their

superficiality and underscores the selective pressure on Dicer to produce siRNAs

within a narrow size range.

The increased activity and ease of expression of the DPR unit is particularly

interesting from a technological standpoint. Although commercially available Dicer

kits and protocols with proprietary modified Dicer variants that cleave dsRNA more

efficiently are available, the DPR unit is a promising avenue for labs more inclined

towards to take a do-it-yourself position. That DPR may differ slightly it is ability to

make the first cut on a dsRNA substrate lacking 3‟2nt overhangs is an intriguing

hypothesis suggesting that the ATPase/Helicase domain has been selectively retained

to process long dsRNA which argues for the presence of human endo-siRNAs and a

potentially robust anti-viral component in human RNAi.

Overall, these results argue that many details with regards to Dicer domain function

are still unknown and that subtle effects can be revealed with relatively simple

experiments. However, more experiments are necessary before strong conclusions can

be made. The experiments presented must be repeated and done in conjunction with

many of the proposed studies to uphold any of the aforementioned hypotheses. For

instance, all of the in vitro dicing assays were carried out with substrates that do not

146

resemble miRNA precursors processed by human Dicer in vivo. Thus, many of the

observations and insights into the function of the ATPase/Helicase and DUF283

reported might not have any bearing on biologically relevant Dicer activity.

It is safe to say that the basic mechanism by which Dicer generates siRNAs is well

understood. In moving forward with studying Dicer, the most interesting questions

concern the regulation of global miRNA activity through Dicer. The identification of

numerous binding partners with ties to several different cellular pathways suggests

dynamic modulation of Dicer activity. Many of the ubiquitous and highly expressed

miRNAs target cell cycle genes and numerous growth factors. Thus, cellular growth

rate could be determined in part through the global miRNA concentration which is a

function of Dicer activity. Recent work on regulated 3‟UTR shortening provides

evidence for a similar phenomenon in that proliferative cells tend to dial down miRNA

activity by regulating 3‟-UTR length on a genomic scale [306,307]. In much the same

way, Dicer activity may be tuned as another method for regulating growth. The recent

discovery that some miRNAs are on/off toggled through post-transcriptional

modifications to the miRNA precursor raises the possibility that Dicer activity might

also be influenced by specific associative factors; either the modified precursors

themselves or by precursor-specific protein binding partners. Dicer‟s association with

Argonaute proteins and the RISC complex expand the outcomes of dynamic Dicer

regulation. It is not implausible that how and what RISC targets is influenced by either

Dicer‟s binding partners or even through the kinetics of Dicer processing and transfer

of its products to Argonaute. Likewise, the balance between translational inhibition

and mRNA degradation of RISC targeted mRNAs might be influenced upon Dicer

kinetics and the binding partners associated with RISC through Dicer.

147

Systematic Identification of miRNA Targets and Steps in Gene

Expression Regulated by miRNAs

We have presented a successful method for the identification of miRNA targets by

virtue of their association with the RISC complex and specifically with the Argonaute

proteins. One drawback in using microarrays to identify the mRNAs bound to RISC is

the probe limitation of the array in use. To some extent, The HEEBO arrays we

employed discern between transcript variants, but the build is by no means 100%

accurate or exhaustive. Likewise, 3‟-UTR sequences used in our analysis are the

longest retrievable database 3‟UTR entries rather than representative of the actual

sequences bound. This is a weakness because 3‟UTR length varies between cell types

and cell states [306,307]. Although high density tiling arrays are available, next

generation sequencing technology provides the strongest option in terms of an

unbiased approach for the purposes of accurate identification of bound RNAs. Deep

sequencing would also aid in the discovery of novel Ago associated small RNAs (like

endo-siRNAs) if the siRNA generating locus of origin is not included in most array

builds.

Excitingly, one group has already taken the next step and combined Ago IPs with an

RNase step followed by deep sequencing of Ago protected sequence fragments to

hone in on the actual miRNA binding sites in a methodology referred to as high-

throughput sequencing of RNA isolated by crosslinking immunoprecipitation (HITS-

CLIP)[308,309]. Basically, prior to the IP, Ago bound transcripts are cross linked, and

then whittled down by a controlled RNase step leaving only sequence protected by its

physical interaction with Ago and minimal flanking sequence. A post IP size selection

ensures that only fragments of certain sized are used for the production of a

sequencing library. Post sequencing, the small fragments are mapped back to the

genome for miRNA/mRNA identification and binding site mapping. Although in a

few cases our data can be used to infer the seed match(es) important in recruiting an

mRNA to RISC, it does not provide information on the bona fide binding sites. High

resolution miRNA binding site mapping is crucial for teasing out the connections

148

between miRNAs and the mRNAs they recruit to Argonaute from the morass of RISC

associated RNAs one finds in a basic Ago IP snapshot. As we demonstrated, specific

miRNA targets can be identified through the comparison of Ago IP profiles from

naïve cells to cells transfected with a specific miRNA. Alternatively, Ago IPs can be

used to identify miRNAs recruited to RISC by specific gene expression programs

rather than specific miRNAs. Ago IPs preformed along a time course after the

initiation of a developmental program or cell fate decision (stem cell differentiation)

could be used to identify miRNAs and mRNAs important in the engagement of a new

expression program. In this way, Ago IPs can be used to measure the emergence or

disappearance of specific miRNAs and corresponding changes in enrichment of

cognate mRNAs when compared to the null state.

Although this type of analysis bypasses the absolute necessity for discernment of

actual binding sites to explain the recruitment of mRNAs, it would probably require >

~5-fold change in the abundance of a specific miRNA to deconvolute most

miRNA:mRNA target pairs given that most mRNAs have multiple seed matches to

numerous miRNAs. Furthermore, miRNAs expressed together in coherent expression

often have overlapping target sets. We have moved forward with Ago IPs to quantify

changes in RISC recruitment along time courses in stimulated Jurkat cells and in the

macrophage-like HL60 cells induced to differentiate into neutrophil like cells with the

goal of adapting our original IP method into a technology similar to Ago HITS-CLIP.

High resolution miRNA:mRNA binding site mapping analysis preformed in tandem

with gene expression and translational profiling along developmental time courses is

the eventual goal. Indeed, similar experiments are becoming more common in the field

and promise to revolutionize our understanding of miRNA

function[309,310,311,312,313,314].

miRNAs have been characterized largely as micro managers of gene expression

important for clearing out mRNAs no longer required by the gene expression

program, counteracting spurious transcription for maintenance of transcriptome

integrity, and by fine tuning expression levels of expressed genes. Nevertheless,

149

growing evidence is building a case that miRNAs are crucial for driving cell fate

decisions. Recent work in human neurons reveals that proper expression of miR-124

and miR-9* is essential to post mitotic neural development[315]. Precocious

expression of miR-124 and miR-9*inhibits proliferation of the neural progenitors.

Zhang and colleagues systematically identified temporally regulated larval stage-

specific miRNA sets using RISC IPs in C. elegans[314]. They found that these

miRNAs coordinate gene expression programs through targeting of signaling

molecules in a larval stage specific manner. In addition, miRNA expression profiling

suggests that many miRNAs exhibit dynamic responses to cell cycle phase[316].

Debating whether miRNA expression drives cell fate decisions rather than maintains

them is largely a semantic argument, as it well established that disrupting miRNA

expression can results in devastating phenotypic consequences. The classification, to a

large extent, hinges on whether or not experimentally induced miRNA expression can

be sufficient to force cells into a specific differentiation trajectory.

Although in our experiments miRNA expression mirrored miRNA RISC

incorporation, it is likely that cells regulate miRNAs inclusion into functional RISC.

Several studies provide evidence that miRNAs can be localized through sequence

localization motifs and by RBPs associated with target mRNAs[214,317]. It would be

interesting to perform sequential IPs of Ago followed by other interesting RISC

cofactors such as FMRP or HUR to identify subsets of miRNAs:mRNAs that have

specialized function or activity. Although biologically dominant mechanisms and

modes of function provide useful theoretical frameworks from which to proceed from,

it is the anomalous examples that sometimes provide the most interesting or

unexpected biology.

The next step in understanding the mechanisms by which miRNAs regulate gene

expression will be to determine if the concordant relationship between mRNA

abundance and translation rate we reported is potentially a function of our usage of an

exogenous miRNA not native to HEK293T cells. Although the targets we identified

closely parallel those reported in mouse brain, and thus are likely bona fide biological

150

targets, HEK cells may not be equipped to regulate these targeted mRNAs the same

way they would in neuronal cells[309]. Brain specific RBPs may bind many of these

messages and regulate the balance of miRNA mediated repression in favor of

translational inhibition or mRNA degradation. It is also possible that in a neuronal

context brain specific RBPs might protect or sequester a subset of miR-124 targets.

We are currently knocking down endogenous let-7 in an effort to measure all the same

parameters of gene expression and determine if there is a strong correlation between

let-7 specific de-repression of mRNA stability and translation. We are also attempting

to incorporate methodologies similar to those described by Ingolia e al to generate

high resolution ribosome positional data to test if there are interesting patterns of

ribosome stalling in miRNA repressed mRNAs [284]. The method is similar to the

Ago HITS-CLP, in that ribosomes are used as an RNA protective agent in a controlled

RNase reaction. Monosomes are then isolated and the associated small RNA

fragments are used for deep sequencing library generation.

Another unanswered question regarding the concordant relationship between miRNA-

mediated reductions in mRNA abundance and translation surrounds the nature of the

functional linkage between the two regulatory fates. Two separate models can account

for this relationship. One possibility is that miRNAs impose an initial translational

repression that leads to deadenylation and subsequent degradation and thus accounting

for the high correlation between these two parameters. On the other hand, miRNA

induced changes in mRNA stability and translation rate may be mechanistically

distinct but initiated by the same RISC cofactors. The situation is even more tangled

when considering that even if the second model were correct, reduced translation

increases degradation and vice versa as translation and degradation are inexorably

linked in a miRNA independent manner. We know that the correlation we reported

between reduced translation and mRNA abundance is much stronger than for non-

targets, and is thus primarily miRNA dependent. However the linkage serves to

complicate attempts to discern between the two models.

151

In part, the question is best framed as one of miRNA induced deadenylation and

translational repression. Recent work from Chen and colleagues carefully measured

the kinetics of miRNA induced deadenylation and degradation[189]. They found that

miRNA specific degradation requires the CCR4/CAF1 deadenylase complex and is

faster than nonsense mediated decay and ARE mediated decay suggesting that miRNA

mediated decay is an intrinsic property of functional RISC[189]. Importantly, the

authors argue that this directed deadenylation is not driven by specific recruitment of

the deadenylating and decapping complexes as there is a notable lack of protein-

protein interaction data between RISC complexes and deadenylation

machinery[191,264,318,319]. Instead the argument is made, based on two separate

studies that the RISC cofactor GW182 interacts with Poly A binding protein (PABP)

to disrupt and inhibit mRNA circularization which has the effect of increasing a

targeted mRNAs susceptibility to engagement with the deadenylation

complex[184,189].

Two recent studies measuring miRNA specific translation repression and

deadenylation found contrasting results[184,320]. In a cell free in vitro system Fabien

et al found that an initial translation repression occurs within 15 minutes of an mRNAs

exposure to a targeting miRNA followed by a slow increase in deadenylation leading

to decay [184]. The authors also found compelling evidence for a dependence on

GW182 and PABP for efficient repression. This study provides strong biochemical

support for the sequential model of miRNA mediated translational inhibition and

mRNA decay. However, the contrasting work from Beilharz et al reports that GW182

dependant polyA tail shortening induced by miRNA targeting precedes translational

repression and can occur in the absence of active translation albeit less

efficiently[320]. It is unclear why these two groups came to such conflicting

conclusions, and this chicken or the egg issue will remain unresolved until more data

is acquired. One possibility for the confusion is that perhaps miRNA targeted mRNAs

can re-enter the translation pool if they are not rapidly destroyed, thus masking

measurable effects on translation in one system versus the other. It is clear however,

152

that miRNAs can directly induce translational repression that is independent from

deadenylation (decay) of their targets and vice versa.

153

Appendix

Text S1. miRNA-effector Complexes Appear to Nonspecifically Bind

Streptavidin Coated Dynal Beads.

One explanation for the high correlation (r=0.6) we observed between the enrichment

profiles of the Ago IPs and negative controls is that the “background” signature of the

negative control IPs was in part driven by non-specific binding of miRNA effector

proteins, such as the Agos, to the beads in the absence of an Ago specific antibody.

We reasoned that if the beads were nonspecifically enriching miRNA effector

proteins, there would be enrichment for mRNAs that are recruited to Agos by miR-

124 in miR-124 negative control IPs compared to negative control IPs from mock

transfected cells.

To test this hypothesis, we determined if mRNAs most enriched in miR-124 negative

control IPs compared to negative control IPs from mock transfected cells were more

likely to contain seed matches to miR-124 in their 3‟UTRs than expected by chance.

First, we tested if any RNAs were specifically recruited to the beads by miR-124 using

SAM to compare negative control IPs from mock versus miR-124 transfected cells.

Only two of the three samples from mock transfected cells yielded microarray data of

high enough quality for use in this analysis. No mRNAs were significantly enriched in

the miR-124 negative control IPs at a 1%, 10%, or 50% FDR. However, hierarchical

cluster analysis of the 500 most enriched sequences from SAM analysis (without

recourse to statistical significance) segregated the miR-124 negative control IPs from

the control IPs with mock transfected cells (Figure S1A). This population was slightly

enriched for mRNAs that contained miR-124 seed matches in their 3‟-UTRs (Figure

S1B). These results suggest that the enrichment profiles from the negative control IPs

are in part, generated from binding of miRNA effector proteins to the beads alone. The

nonspecific binding is relatively weak as evidenced by the low amount of RNA and

154

protein isolated from the beads (10x less than Ago IPs) and the fact that no individual

mRNAs were enriched with high-statistical confidence in miR-124 negative control

IPs compared to negative control IPs from mock transfected cells. Perhaps recruitment

of miRNA effector proteins to the beads is driven by the association of Ago proteins

with chaperones, such as Hsp90 [292,319, unpublished data].

155

Text S2. Enrichment of Seed Matches to Highly-expressed miRNAs in

Ago IPs from Mock Transfected Cells.

To test whether association with Ago is largely a reflection of the relative occupancy

of each mRNA with the suite of endogenously expressed miRNAs in HEK293T cells,

we first determined if there were sequence motifs in mRNA untranslated regions that

significantly correlated with Ago IP enrichment, and if so, if these motifs

corresponded to known miRNA seed matches. Such sequence motifs might be

identifiable if one or a few highly expressed miRNAs dominated the IP enrichment.

To search for possible sequence motifs, we used the motif prediction algorithm

FIRE[321].Two motifs, located in 3‟-UTRs, significantly correlated with Ago IP

enrichment; these motifs overlapped and corresponded to seed match sequences of the

very abundant miR-17-5p/20/92/106/591.d and miR-19 families of miRNAs. mRNAs

with seed matches to these miRNAs in good sequence contexts (TargetScan 4.2,

context score < -0.3) were more likely to be enriched in Ago IPs than mRNAs with

seed matches in poor sequence contexts (context score > -0.1) (for miR-17: 2.9 mean

fold-enrichment for mRNAs in good context versus 0.91 mean fold-enrichment for

mRNAs in poor context, p < 10-12

, one-sided Kolmogorov-Smirnov test)[88]. This

result suggests that these abundant miRNAs contribute significantly to recruitment to

RISC and Ago IP enrichment. For a majority of expressed miRNAs, however, we did

not find that predicted targets with seed matches in a good context were more likely to

be enriched than mRNAs with seed matches in a poor context. This negative finding

may result because mRNA enrichment in Ago IPs arises from multiple miRNAs so

that the enrichment signal for any specific miRNA is diluted.

156

Text S3. Relationship Between Ribosome Occupancy in Mock-

Transfected Cells and the Change in Ribosome Occupancy Following

Transfection of miR-124.

Twenty-one miR-124 Ago IP targets appeared to increase at least 20% in ribosome

occupancy due to the presence of miR-124. Strikingly, a large fraction of mRNAs that

increased at least 20% in ribosome occupancy encode transcription factors (11/21, p <

10-4

).

We asked if miR-124 targets that increase at least 20% in ribosome occupancy behave

anomalously in other ways, which might provide some insight into their apparent

increase in occupancy. We compared mRNAs that increase at least 20% in ribosome

occupancy to the other miR-124 Ago IP targets with regards to several factors. While

there was no difference in terms of the fraction of each group that contained seed

matches to miR-124 or their ribosome density, targets that increased in ribosome

occupancy were less likely to decrease in abundance (13% vs 33%) and had much

lower ribosome occupancy in untreated cells than the other miR-124 targets (49% vs

87%).

The fact that the mRNA targets that appeared to increase in ribosome occupancy

tended to have low occupancy in mock transfected cells prompted us to ask if this was

a general phenomenon. We found that across all mRNAs there was a strong negative

correlation between the miR-124-induced change in ribosome occupancy of a gene‟s

mRNAs and its ribosome occupancy in mock transfected cells (Spearman rank

correlation = -0.78) (Figure S4B). This relationship could, in principle, be due to

errors in our microarray measurements, which would tend to have a larger effect on

the estimated ribosome occupancy as ribosome occupancy decreases. If this were the

case we would expect significant correlations between the change in ribosome

occupancy and ribosome occupancy between biological replicate experiments.

However, the relationship between the change in ribosome occupancy as a function of

157

ribosome occupancy is much weaker between biological replicates (r = -0.07, 0.24 and

0.02 for mock replicates and -0.11 for miR-124 replicate) than between mock and

miR-124 experiments (r = -0.55, -0.45, -0.66, -0.60, -0.74 and -0.77 between mock

and miR-124 experiments), which suggests it may actually be due to bona fide

biological differences between mock and miR-124 transfected cells. Indeed, mRNAs

that have low initial ribosome occupancy have more potential to increase in ribosome

occupancy between experimental conditions (Figure S4B, gray curve). For instance, a

gene whose mRNAs are 10% occupied can increase up to 10-fold, while a gene whose

mRNAs are 90% occupied increase no more than 1.1 fold.

While the origin of the relationship between changes in ribosome occupancy and

starting ribosome occupancy might be biological, we still wished to investigate the

effect of this potential artifact on our results. We attempted to remove the relationship

between changes in ribosome occupancy and starting ribosome occupancy by fitting a

locally weighted scatterplot smoothing (Lowess) function to the scatterplot between

mock ribosome occupancy and change in ribosome occupancy and subtracting the

fitted values (Figure S4B – green circles)[322]. As expected, after this transformation,

there was no longer a correlation between ribosome occupancy in mock experiments

and the miR-124-induced change in ribosome occupancy (Figure S4C). The overall

effects on ribosome occupancy and translation were very similar between the

renormalized data and raw data (mean difference in occupancy between targets and

nontargets with Lowess normalization = 4%, and without Lowess normalization = 4%)

and the correlation between changes in translation and abundance for miR-124 Ago IP

targets were very similar (Lowess-normalized = 0.59, without Lowess normalization =

0.60) (Figure S4D and S4E). However, the group of mRNAs that, prior to this

transformation, had appeared to increase in translation in the miR-124 transfected cells

no longer appeared to do so after the Lowess normalization. Following the Lowess

normalization, the IP targets that previously appeared to increase in occupancy by at

least 20% had, on average, no change in occupancy, there were no IP targets with >

13% apparent increase in occupancy, and there was no functional bias among mRNAs

that appeared to increase in occupancy.

158

Text S4. Evaluation of the Significance of the Correlation between

Changes in mRNA Abundance and Translation of miR-124 Ago IP

targets Following Transfection with miR-124.

We evaluate the significance of the correlation between expression changes and

translation changes in several ways. First, we asked from a purely statistical

perspective, what was the likelihood of observing the correlation by chance, given the

values we had. To test this, we calculated the correlation between expression and

translation after resampling the expression data. We did this 100,000 times.

According the normal distribution function, the likelihood of getting a correlation of

0.60 by chance is 10-45

(Figure S5A). Second, we tested if the correlation persisted in

nontarget mRNAs. We found that the Pearson correlation between expression

changes and translation changes for mRNAs that are not IP targets was 0.11. We

ranked mRNAs by IP enrichment and plotted the correlation between changes in

expression and translation as a moving window, and found that the correlation

monotonically falls off to baseline level after the several hundred most enriched

mRNAs, indicating the correlation is specific to mRNAs most enriched in Ago IPs due

to the presence of miR-124 (Figure S5B). Third, we tested if the correlation was

specific to miR-124 targets, or if the observed relationship between mRNA abundance

and translation was a general phenomenon. It is well documented that changes in

abundance and translation tend to correlate and so it may not be specific to miR-124

response, but rather reflect a common relationship between effects on translation and

mRNA abundance in this system

[215,231,232,233,234,294,295,296,298,299,300,301,302]. To estimate the difference

in the strength of the correlation between translation and mRNA abundance we chose

nontarget mRNAs that decreased at the mRNA level similarly to miR-124 target

mRNAs. We had to focus on miR-124 targets that changed less than 40% in mRNA

abundance as there were not any nontargets that decreased more than 40%. We

compared the miR-124 IP targets that decreased less than 40% in mRNA abundance to

10,000 random sets of nontargets of the same size with similar distributions of

expression changes (t-test, p > 0.001). The average correlation for the 10,000 sets of

159

nontargets was 0.14, whereas the correlation between translation and expression for

miR-124 IP targets changing in mRNA abundance <40% was 0.30 (1/10,000

permuted nontargets sets had a correlation greater than the IP targets – we estimated

the actual p-value using normal distribution function to be < 10-5

) (Figure S5C). We

also observed a modest, but significant, positive correlation between changes in

expression and translation for mRNAs whose abundance increases (0.19 for 662

mRNAs that increase at least 25%.) These data suggest changes in abundance and

translation are generally correlated under these growth conditions, but tend to be more

so for miR-124 targets.

160

References

1. Ghaemmaghami S, Huh WK, Bower K, Howson RW, Belle A, et al. (2003) Global

analysis of protein expression in yeast. Nature 425: 737-741.

2. Gry M, Rimini R, Stromberg S, Asplund A, Ponten F, et al. (2009) Correlations

between RNA and protein expression profiles in 23 human cell lines. BMC

Genomics 10: 365.

3. Gygi SP, Rochon Y, Franza BR, Aebersold R (1999) Correlation between protein

and mRNA abundance in yeast. Mol Cell Biol 19: 1720-1730.

4. Kuhn KM, DeRisi JL, Brown PO, Sarnow P (2001) Global and specific

translational regulation in the genomic response of Saccharomyces cerevisiae

to a rapid transfer from a fermentable to a nonfermentable carbon source. Mol

Cell Biol 21: 916-927.

5. Lu P, Vogel C, Wang R, Yao X, Marcotte EM (2007) Absolute protein expression

profiling estimates the relative contributions of transcriptional and translational

regulation. Nat Biotechnol 25: 117-124.

6. Newman JR, Ghaemmaghami S, Ihmels J, Breslow DK, Noble M, et al. (2006)

Single-cell proteomic analysis of S. cerevisiae reveals the architecture of

biological noise. Nature 441: 840-846.

7. Shankavaram UT, Reinhold WC, Nishizuka S, Major S, Morita D, et al. (2007)

Transcript and protein expression profiles of the NCI-60 cancer cell panel: an

integromic microarray study. Mol Cancer Ther 6: 820-832.

8. Wightman B, Ha I, Ruvkun G (1993) Posttranscriptional regulation of the

heterochronic gene lin-14 by lin-4 mediates temporal pattern formation in C.

elegans. Cell 75: 855-862.

9. Lee RC, Feinbaum RL, Ambros V (1993) The C. elegans heterochronic gene lin-4

encodes small RNAs with antisense complementarity to lin-14. Cell 75: 843-

854.

161

10. Lau NC, Lim LP, Weinstein EG, Bartel DP (2001) An abundant class of tiny

RNAs with probable regulatory roles in Caenorhabditis elegans. Science 294:

858-862.

11. Lagos-Quintana M, Rauhut R, Lendeckel W, Tuschl T (2001) Identification of

novel genes coding for small expressed RNAs. Science 294: 853-858.

12. Lee RC, Ambros V (2001) An extensive class of small RNAs in Caenorhabditis

elegans. Science 294: 862-864.

13. van der Krol AR, Mur LA, de Lange P, Mol JN, Stuitje AR (1990) Inhibition of

flower pigmentation by antisense CHS genes: promoter and minimal sequence

requirements for the antisense effect. Plant Mol Biol 14: 457-466.

14. Napoli C, Lemieux C, Jorgensen R (1990) Introduction of a Chimeric Chalcone

Synthase Gene into Petunia Results in Reversible Co-Suppression of

Homologous Genes in trans. Plant Cell 2: 279-289.

15. Anandalakshmi R, Pruss GJ, Ge X, Marathe R, Mallory AC, et al. (1998) A viral

suppressor of gene silencing in plants. Proc Natl Acad Sci U S A 95: 13079-

13084.

16. Ruiz MT, Voinnet O, Baulcombe DC (1998) Initiation and maintenance of virus-

induced gene silencing. Plant Cell 10: 937-946.

17. Angell SM, Baulcombe DC (1997) Consistent gene silencing in transgenic plants

expressing a replicating potato virus X RNA. Embo J 16: 3675-3684.

18. Dougherty WG, Lindbo JA, Smith HA, Parks TD, Swaney S, et al. (1994) RNA-

mediated virus resistance in transgenic plants: exploitation of a cellular

pathway possibly involved in RNA degradation. Mol Plant Microbe Interact 7:

544-552.

19. Kumagai MH, Donson J, della-Cioppa G, Harvey D, Hanley K, et al. (1995)

Cytoplasmic inhibition of carotenoid biosynthesis with virus-derived RNA.

Proc Natl Acad Sci U S A 92: 1679-1683.

20. Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, et al. (1998) Potent and

specific genetic interference by double-stranded RNA in Caenorhabditis

elegans. Nature 391: 806-811.

162

21. Carrington JC, Ambros V (2003) Role of microRNAs in plant and animal

development. Science 301: 336-338.

22. Provost P, Silverstein RA, Dishart D, Walfridsson J, Djupedal I, et al. (2002)

Dicer is required for chromosome segregation and gene silencing in fission

yeast cells. Proc Natl Acad Sci U S A 99: 16648-16653.

23. Verdel A, Jia S, Gerber S, Sugiyama T, Gygi S, et al. (2004) RNAi-mediated

targeting of heterochromatin by the RITS complex. Science 303: 672-676.

24. Xie Z, Johansen LK, Gustafson AM, Kasschau KD, Lellis AD, et al. (2004)

Genetic and functional diversification of small RNA pathways in plants. PLoS

Biol 2: E104.

25. Ambros V (2004) The functions of animal microRNAs. Nature 431: 350-355.

26. Bagga S, Bracht J, Hunter S, Massirer K, Holtz J, et al. (2005) Regulation by let-7

and lin-4 miRNAs results in target mRNA degradation. Cell 122: 553-563.

27. Filipowicz W, Bhattacharyya SN, Sonenberg N (2008) Mechanisms of post-

transcriptional regulation by microRNAs: are the answers in sight? Nat Rev

Genet 9: 102-114.

28. Humphreys DT, Westman BJ, Martin DI, Preiss T (2005) MicroRNAs control

translation initiation by inhibiting eukaryotic initiation factor 4E/cap and

poly(A) tail function. Proc Natl Acad Sci U S A 102: 16961-16966.

29. Lim LP, Lau NC, Garrett-Engele P, Grimson A, Schelter JM, et al. (2005)

Microarray analysis shows that some microRNAs downregulate large numbers

of target mRNAs. Nature 433: 769-773.

30. Olsen PH, Ambros V (1999) The lin-4 regulatory RNA controls developmental

timing in Caenorhabditis elegans by blocking LIN-14 protein synthesis after

the initiation of translation. Dev Biol 216: 671-680.

31. Valencia-Sanchez MA, Liu J, Hannon GJ, Parker R (2006) Control of translation

and mRNA degradation by miRNAs and siRNAs. Genes Dev 20: 515-524.

32. Wu L, Belasco JG (2008) Let me count the ways: mechanisms of gene regulation

by miRNAs and siRNAs. Mol Cell 29: 1-7.

33. Wu L, Fan J, Belasco JG (2006) MicroRNAs direct rapid deadenylation of mRNA.

Proc Natl Acad Sci U S A 103: 4034-4039.

163

34. Drinnenberg IA, Weinberg DE, Xie KT, Mower JP, Wolfe KH, et al. (2009) RNAi

in Budding Yeast. Science.

35. Hannon GJ (2002) RNA interference. Nature 418: 244-251.

36. Bernstein E, Caudy AA, Hammond SM, Hannon GJ (2001) Role for a bidentate

ribonuclease in the initiation step of RNA interference. Nature 409: 363-366.

37. Hutvagner G, McLachlan J, Pasquinelli AE, Balint E, Tuschl T, et al. (2001) A

cellular function for the RNA-interference enzyme Dicer in the maturation of

the let-7 small temporal RNA. Science 293: 834-838.

38. Ketting RF, Fischer SE, Bernstein E, Sijen T, Hannon GJ, et al. (2001) Dicer

functions in RNA interference and in synthesis of small RNA involved in

developmental timing in C. elegans. Genes Dev 15: 2654-2659.

39. Golden TA, Schauer SE, Lang JD, Pien S, Mushegian AR, et al. (2002) SHORT

INTEGUMENTS1/SUSPENSOR1/CARPEL FACTORY, a Dicer homolog, is

a maternal effect gene required for embryo development in Arabidopsis. Plant

Physiol 130: 808-822.

40. Hutvagner G, Zamore PD (2002) A microRNA in a multiple-turnover RNAi

enzyme complex. Science 297: 2056-2060.

41. Provost P, Dishart D, Doucet J, Frendewey D, Samuelsson B, et al. (2002)

Ribonuclease activity and RNA binding of recombinant human Dicer. Embo J

21: 5864-5874.

42. Zhang H, Kolb FA, Brondani V, Billy E, Filipowicz W (2002) Human Dicer

preferentially cleaves dsRNAs at their termini without a requirement for ATP.

Embo J 21: 5875-5885.

43. Elbashir SM, Harborth J, Lendeckel W, Yalcin A, Weber K, et al. (2001) Duplexes

of 21-nucleotide RNAs mediate RNA interference in cultured mammalian

cells. Nature 411: 494-498.

44. Elbashir SM, Lendeckel W, Tuschl T (2001) RNA interference is mediated by 21-

and 22-nucleotide RNAs. Genes Dev 15: 188-200.

45. Elbashir SM, Martinez J, Patkaniowska A, Lendeckel W, Tuschl T (2001)

Functional anatomy of siRNAs for mediating efficient RNAi in Drosophila

melanogaster embryo lysate. Embo J 20: 6877-6888.

164

46. Schwarz DS, Hutvagner G, Haley B, Zamore PD (2002) Evidence that siRNAs

function as guides, not primers, in the Drosophila and human RNAi pathways.

Mol Cell 10: 537-548.

47. Hutvagner G, Zamore PD (2002) RNAi: nature abhors a double-strand. Curr Opin

Genet Dev 12: 225-232.

48. Martinez J, Patkaniowska A, Urlaub H, Luhrmann R, Tuschl T (2002) Single-

stranded antisense siRNAs guide target RNA cleavage in RNAi. Cell 110: 563-

574.

49. Hammond SM, Boettcher S, Caudy AA, Kobayashi R, Hannon GJ (2001)

Argonaute2, a link between genetic and biochemical analyses of RNAi.

Science 293: 1146-1150.

50. Schwarz DS, Hutvagner G, Du T, Xu Z, Aronin N, et al. (2003) Asymmetry in the

assembly of the RNAi enzyme complex. Cell 115: 199-208.

51. Song JJ, Liu J, Tolia NH, Schneiderman J, Smith SK, et al. (2003) The crystal

structure of the Argonaute2 PAZ domain reveals an RNA binding motif in

RNAi effector complexes. Nat Struct Biol 10: 1026-1032.

52. Carmell MA, Xuan Z, Zhang MQ, Hannon GJ (2002) The Argonaute family:

tentacles that reach into RNAi, developmental control, stem cell maintenance,

and tumorigenesis. Genes Dev 16: 2733-2742.

53. Hutvagner G, Simard MJ, Mello CC, Zamore PD (2004) Sequence-specific

inhibition of small RNA function. PLoS Biol 2: E98.

54. Liu J, Carmell MA, Rivas FV, Marsden CG, Thomson JM, et al. (2004)

Argonaute2 Is the Catalytic Engine of Mammalian RNAi. Science.

55. Behm-Ansmant I, Rehwinkel J, Doerks T, Stark A, Bork P, et al. (2006) mRNA

degradation by miRNAs and GW182 requires both CCR4:NOT deadenylase

and DCP1:DCP2 decapping complexes. Genes Dev 20: 1885-1898.

56. Rehwinkel J, Behm-Ansmant I, Gatfield D, Izaurralde E (2005) A crucial role for

GW182 and the DCP1:DCP2 decapping complex in miRNA-mediated gene

silencing. Rna 11: 1640-1647.

57. Eulalio A, Huntzinger E, Nishihara T, Rehwinkel J, Fauser M, et al. (2009)

Deadenylation is a widespread effect of miRNA regulation. Rna 15: 21-32.

165

58. Doench JG, Petersen CP, Sharp PA (2003) siRNAs can function as miRNAs.

Genes Dev 17: 438-442.

59. Bartel DP (2004) MicroRNAs: genomics, biogenesis, mechanism, and function.

Cell 116: 281-297.

60. Doench JG, Sharp PA (2004) Specificity of microRNA target selection in

translational repression. Genes Dev 18: 504-511.

61. Wakiyama M, Takimoto K, Ohara O, Yokoyama S (2007) Let-7 microRNA-

mediated mRNA deadenylation and translational repression in a mammalian

cell-free system. Genes Dev 21: 1857-1862.

62. Zeng Y, Yi R, Cullen BR (2003) MicroRNAs and small interfering RNAs can

inhibit mRNA expression by similar mechanisms. Proc Natl Acad Sci U S A

100: 9779-9784.

63. Meister G, Landthaler M, Patkaniowska A, Dorsett Y, Teng G, et al. (2004)

Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs.

Mol Cell 15: 185-197.

64. Carthew RW, Sontheimer EJ (2009) Origins and Mechanisms of miRNAs and

siRNAs. Cell 136: 642-655.

65. Motamedi MR, Verdel A, Colmenares SU, Gerber SA, Gygi SP, et al. (2004) Two

RNAi complexes, RITS and RDRC, physically interact and localize to

noncoding centromeric RNAs. Cell 119: 789-802.

66. Noma K, Sugiyama T, Cam H, Verdel A, Zofall M, et al. (2004) RITS acts in cis

to promote RNA interference-mediated transcriptional and post-transcriptional

silencing. Nat Genet 36: 1174-1180.

67. Ekwall K (2004) The RITS complex-A direct link between small RNA and

heterochromatin. Mol Cell 13: 304-305.

68. Bagasra O, Prilliman KR (2004) RNA interference: the molecular immune system.

J Mol Histol 35: 545-553.

69. Okamura K, Lai EC (2008) Endogenous small interfering RNAs in animals. Nat

Rev Mol Cell Biol 9: 673-678.

70. Nilsen TW (2008) Endo-siRNAs: yet another layer of complexity in RNA

silencing. Nat Struct Mol Biol 15: 546-548.

166

71. Ghildiyal M, Seitz H, Horwich MD, Li C, Du T, et al. (2008) Endogenous siRNAs

derived from transposons and mRNAs in Drosophila somatic cells. Science

320: 1077-1081.

72. Lee RC, Hammell CM, Ambros V (2006) Interacting endogenous and exogenous

RNAi pathways in Caenorhabditis elegans. Rna 12: 589-597.

73. Chung WJ, Okamura K, Martin R, Lai EC (2008) Endogenous RNA interference

provides a somatic defense against Drosophila transposons. Curr Biol 18: 795-

802.

74. Deddouche S, Matt N, Budd A, Mueller S, Kemp C, et al. (2008) The DExD/H-

box helicase Dicer-2 mediates the induction of antiviral activity in drosophila.

Nat Immunol 9: 1425-1432.

75. Smardon A, Spoerke JM, Stacey SC, Klein ME, Mackin N, et al. (2000) EGO-1 is

related to RNA-directed RNA polymerase and functions in germ-line

development and RNA interference in C. elegans. Curr Biol 10: 169-178.

76. Sijen T, Fleenor J, Simmer F, Thijssen KL, Parrish S, et al. (2001) On the role of

RNA amplification in dsRNA-triggered gene silencing. Cell 107: 465-476.

77. Alder MN, Dames S, Gaudet J, Mango SE (2003) Gene silencing in

Caenorhabditis elegans by transitive RNA interference. Rna 9: 25-32.

78. Lipardi C, Paterson BM (2009) Identification of an RNA-dependent RNA

polymerase in Drosophila involved in RNAi and transposon suppression. Proc

Natl Acad Sci U S A 106: 15645-15650.

79. Stark JJ, Dillman RO, Schulof R, Wiemann MC, Barth NM, et al. (1998)

Interferon-alpha and chemohormonal therapy for patients with advanced

melanoma: final results of a phase I-II study of the Cancer Biotherapy

Research Group and the Mid-Atlantic Oncology Program. Cancer 82: 1677-

1681.

80. Hannon GJ, Conklin DS (2004) RNA interference by short hairpin RNAs

expressed in vertebrate cells. Methods Mol Biol 257: 255-266.

81. Myers JW, Jones JT, Meyer T, Ferrell JE, Jr. (2003) Recombinant Dicer efficiently

converts large dsRNAs into siRNAs suitable for gene silencing. Nat

Biotechnol 21: 324-328.

167

82. Paddison PJ, Caudy AA, Sachidanandam R, Hannon GJ (2004) Short hairpin

activated gene silencing in mammalian cells. Methods Mol Biol 265: 85-100.

83. Lewis BP, Burge CB, Bartel DP (2005) Conserved seed pairing, often flanked by

adenosines, indicates that thousands of human genes are microRNA targets.

Cell 120: 15-20.

84. Friedman RC, Farh KK, Burge CB, Bartel D (2008) Most mammalian mRNAs are

conserved targets of microRNAs. Genome Res.

85. Grimson A, Srivastava M, Fahey B, Woodcroft BJ, Chiang HR, et al. (2008) Early

origins and evolution of microRNAs and Piwi-interacting RNAs in animals.

Nature 455: 1193-1197.

86. Farh KK, Grimson A, Jan C, Lewis BP, Johnston WK, et al. (2005) The

widespread impact of mammalian MicroRNAs on mRNA repression and

evolution. Science 310: 1817-1821.

87. Bartel DP (2009) MicroRNAs: target recognition and regulatory functions. Cell

136: 215-233.

88. Grimson A, Farh KK, Johnston WK, Garrett-Engele P, Lim LP, et al. (2007)

MicroRNA targeting specificity in mammals: determinants beyond seed

pairing. Mol Cell 27: 91-105.

89. Lewis BP, Shih IH, Jones-Rhoades MW, Bartel DP, Burge CB (2003) Prediction

of mammalian microRNA targets. Cell 115: 787-798.

90. Baek D, Villen J, Shin C, Camargo FD, Gygi SP, et al. (2008) The impact of

microRNAs on protein output. Nature 455: 64-71.

91. Selbach M, Schwanhausser B, Thierfelder N, Fang Z, Khanin R, et al. (2008)

Widespread changes in protein synthesis induced by microRNAs. Nature 455:

58-63.

92. Abbott AL, Alvarez-Saavedra E, Miska EA, Lau NC, Bartel DP, et al. (2005) The

let-7 MicroRNA family members mir-48, mir-84, and mir-241 function

together to regulate developmental timing in Caenorhabditis elegans. Dev Cell

9: 403-414.

93. Alvarez-Garcia I, Miska EA (2005) MicroRNA functions in animal development

and human disease. Development 132: 4653-4662.

168

94. Hatfield SD, Shcherbata HR, Fischer KA, Nakahara K, Carthew RW, et al. (2005)

Stem cell division is regulated by the microRNA pathway. Nature 435: 974-

978.

95. Kloosterman WP, Plasterk RH (2006) The diverse functions of microRNAs in

animal development and disease. Dev Cell 11: 441-450.

96. Linsley PS, Schelter J, Burchard J, Kibukawa M, Martin MM, et al. (2007)

Transcripts targeted by the microRNA-16 family cooperatively regulate cell

cycle progression. Mol Cell Biol 27: 2240-2252.

97. Lu J, Getz G, Miska EA, Alvarez-Saavedra E, Lamb J, et al. (2005) MicroRNA

expression profiles classify human cancers. Nature 435: 834-838.

98. Reinhart BJ, Slack FJ, Basson M, Pasquinelli AE, Bettinger JC, et al. (2000) The

21-nucleotide let-7 RNA regulates developmental timing in Caenorhabditis

elegans. Nature 403: 901-906.

99. Jackson RJ, Standart N (2007) How do microRNAs regulate gene expression? Sci

STKE 2007: re1.

100. Esau C, Davis S, Murray SF, Yu XX, Pandey SK, et al. (2006) miR-122

regulation of lipid metabolism revealed by in vivo antisense targeting. Cell

Metab 3: 87-98.

101. Eulalio A, Rehwinkel J, Stricker M, Huntzinger E, Yang SF, et al. (2007) Target-

specific requirements for enhancers of decapping in miRNA-mediated gene

silencing. Genes Dev 21: 2558-2570.

102. Giraldez AJ, Mishima Y, Rihel J, Grocock RJ, Van Dongen S, et al. (2006)

Zebrafish MiR-430 promotes deadenylation and clearance of maternal

mRNAs. Science 312: 75-79.

103. Krutzfeldt J, Rajewsky N, Braich R, Rajeev KG, Tuschl T, et al. (2005) Silencing

of microRNAs in vivo with 'antagomirs'. Nature 438: 685-689.

104. Rehwinkel J, Natalin P, Stark A, Brennecke J, Cohen SM, et al. (2006) Genome-

wide analysis of mRNAs regulated by Drosha and Argonaute proteins in

Drosophila melanogaster. Mol Cell Biol 26: 2965-2975.

169

105. Schmitter D, Filkowski J, Sewer A, Pillai RS, Oakeley EJ, et al. (2006) Effects of

Dicer and Argonaute down-regulation on mRNA levels in human HEK293

cells. Nucleic Acids Res 34: 4801-4815.

106. Wu L, Belasco JG (2005) Micro-RNA regulation of the mammalian lin-28 gene

during neuronal differentiation of embryonal carcinoma cells. Mol Cell Biol

25: 9198-9208.

107. Nilsen TW (2007) Mechanisms of microRNA-mediated gene regulation in

animal cells. Trends Genet 23: 243-249.

108. Pillai RS, Bhattacharyya SN, Artus CG, Zoller T, Cougot N, et al. (2005)

Inhibition of translational initiation by Let-7 MicroRNA in human cells.

Science 309: 1573-1576.

109. Yekta S, Shih IH, Bartel DP (2004) MicroRNA-directed cleavage of HOXB8

mRNA. Science 304: 594-596.

110. Kozak M (2008) Faulty old ideas about translational regulation paved the way for

current confusion about how microRNAs function. Gene 423: 108-115.

111. Nissan T, Parker R (2008) Computational analysis of miRNA-mediated

repression of translation: implications for models of translation initiation

inhibition. Rna 14: 1480-1491.

112. Shyu AB, Wilkinson MF, van Hoof A (2008) Messenger RNA regulation: to

translate or to degrade. Embo J 27: 471-481.

113. Aravin AA, Hannon GJ, Brennecke J (2007) The Piwi-piRNA pathway provides

an adaptive defense in the transposon arms race. Science 318: 761-764.

114. Song JJ, Smith SK, Hannon GJ, Joshua-Tor L (2004) Crystal Structure of

Argonaute and Its Implications for RISC Slicer Activity. Science.

115. Yan KS, Yan S, Farooq A, Han A, Zeng L, et al. (2003) Structure and conserved

RNA binding of the PAZ domain. Nature 426: 468-474.

116. Lingel A, Simon B, Izaurralde E, Sattler M (2003) Structure and nucleic-acid

binding of the Drosophila Argonaute 2 PAZ domain. Nature 426: 465-469.

117. Ma JB, Ye K, Patel DJ (2004) Structural basis for overhang-specific small

interfering RNA recognition by the PAZ domain. Nature 429: 318-322.

170

118. Macrae IJ, Zhou K, Li F, Repic A, Brooks AN, et al. (2006) Structural basis for

double-stranded RNA processing by Dicer. Science 311: 195-198.

119. MacRae IJ, Zhou K, Doudna JA (2007) Structural determinants of RNA

recognition and cleavage by Dicer. Nat Struct Mol Biol 14: 934-940.

120. Lund E, Guttinger S, Calado A, Dahlberg JE, Kutay U (2004) Nuclear export of

microRNA precursors. Science 303: 95-98.

121. Yi R, Doehle BP, Qin Y, Macara IG, Cullen BR (2005) Overexpression of

exportin 5 enhances RNA interference mediated by short hairpin RNAs and

microRNAs. Rna 11: 220-226.

122. Zeng Y, Yi R, Cullen BR (2005) Recognition and cleavage of primary

microRNA precursors by the nuclear processing enzyme Drosha. Embo J 24:

138-148.

123. Bohnsack MT, Czaplinski K, Gorlich D (2004) Exportin 5 is a RanGTP-

dependent dsRNA-binding protein that mediates nuclear export of pre-

miRNAs. Rna 10: 185-191.

124. Lee Y, Ahn C, Han J, Choi H, Kim J, et al. (2003) The nuclear RNase III Drosha

initiates microRNA processing. Nature 425: 415-419.

125. Ambros V (2001) microRNAs: tiny regulators with great potential. Cell 107:

823-826.

126. Murchison EP, Hannon GJ (2004) miRNAs on the move: miRNA biogenesis and

the RNAi machinery. Curr Opin Cell Biol 16: 223-229.

127. Ying SY, Lin SL (2004) Intron-derived microRNAs--fine tuning of gene

functions. Gene 342: 25-28.

128. Chan SP, Slack FJ (2007) And now introducing mammalian mirtrons. Dev Cell

13: 605-607.

129. Berezikov E, Chung WJ, Willis J, Cuppen E, Lai EC (2007) Mammalian mirtron

genes. Mol Cell 28: 328-336.

130. Okamura K, Hagen JW, Duan H, Tyler DM, Lai EC (2007) The mirtron pathway

generates microRNA-class regulatory RNAs in Drosophila. Cell 130: 89-100.

131. Ruby JG, Jan CH, Bartel DP (2007) Intronic microRNA precursors that bypass

Drosha processing. Nature 448: 83-86.

171

132. Kim VN (2004) MicroRNA precursors in motion: exportin-5 mediates their

nuclear export. Trends Cell Biol 14: 156-159.

133. Lee YS, Nakahara K, Pham JW, Kim K, He Z, et al. (2004) Distinct roles for

Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways. Cell

117: 69-81.

134. Pham JW, Pellino JL, Lee YS, Carthew RW, Sontheimer EJ (2004) A Dicer-2-

dependent 80s complex cleaves targeted mRNAs during RNAi in Drosophila.

Cell 117: 83-94.

135. Nykanen A, Haley B, Zamore PD (2001) ATP requirements and small interfering

RNA structure in the RNA interference pathway. Cell 107: 309-321.

136. Maniataki E, Mourelatos Z (2005) A human, ATP-independent, RISC assembly

machine fueled by pre-miRNA. Genes Dev 19: 2979-2990.

137. Gregory RI, Chendrimada TP, Cooch N, Shiekhattar R (2005) Human RISC

couples microRNA biogenesis and posttranscriptional gene silencing. Cell 123:

631-640.

138. Chendrimada TP, Gregory RI, Kumaraswamy E, Norman J, Cooch N, et al.

(2005) TRBP recruits the Dicer complex to Ago2 for microRNA processing

and gene silencing. Nature 436: 740-744.

139. Myers JW, Ferrell JE (2005) Silencing gene expression with Dicer-generated

siRNA pools. Methods Mol Biol 309: 93-196.

140. Jaskiewicz L, Filipowicz W (2008) Role of Dicer in posttranscriptional RNA

silencing. Curr Top Microbiol Immunol 320: 77-97.

141. Park W, Li J, Song R, Messing J, Chen X (2002) CARPEL FACTORY, a Dicer

homolog, and HEN1, a novel protein, act in microRNA metabolism in

Arabidopsis thaliana. Curr Biol 12: 1484-1495.

142. Tabara H, Yigit E, Siomi H, Mello CC (2002) The dsRNA binding protein RDE-

4 interacts with RDE-1, DCR-1, and a DExH-box helicase to direct RNAi in C.

elegans. Cell 109: 861-871.

143. Liu Q, Rand TA, Kalidas S, Du F, Kim HE, et al. (2003) R2D2, a bridge between

the initiation and effector steps of the Drosophila RNAi pathway. Science 301:

1921-1925.

172

144. Soifer HS, Sano M, Sakurai K, Chomchan P, Saetrom P, et al. (2008) A role for

the Dicer helicase domain in the processing of thermodynamically unstable

hairpin RNAs. Nucleic Acids Res 36: 6511-6522.

145. Dlakic M (2006) DUF283 domain of Dicer proteins has a double-stranded RNA-

binding fold. Bioinformatics 22: 2711-2714.

146. Zhang H, Kolb FA, Jaskiewicz L, Westhof E, Filipowicz W (2004) Single

processing center models for human Dicer and bacterial RNase III. Cell 118:

57-68.

147. Duchaine TF, Wohlschlegel JA, Kennedy S, Bei Y, Conte D, Jr., et al. (2006)

Functional proteomics reveals the biochemical niche of C. elegans DCR-1 in

multiple small-RNA-mediated pathways. Cell 124: 343-354.

148. Caudy AA, Ketting RF, Hammond SM, Denli AM, Bathoorn AM, et al. (2003) A

micrococcal nuclease homologue in RNAi effector complexes. Nature 425:

411-414.

149. Ishizuka A, Siomi MC, Siomi H (2002) A Drosophila fragile X protein interacts

with components of RNAi and ribosomal proteins. Genes Dev 16: 2497-2508.

150. Kotaja N, Bhattacharyya SN, Jaskiewicz L, Kimmins S, Parvinen M, et al. (2006)

The chromatoid body of male germ cells: similarity with processing bodies and

presence of Dicer and microRNA pathway components. Proc Natl Acad Sci U

S A 103: 2647-2652.

151. Tahbaz N, Kolb FA, Zhang H, Jaronczyk K, Filipowicz W, et al. (2004)

Characterization of the interactions between mammalian PAZ PIWI domain

proteins and Dicer. EMBO Rep 5: 189-194.

152. Pare JM, Tahbaz N, Lopez-Orozco J, LaPointe P, Lasko P, et al. (2009) Hsp90

regulates the function of argonaute 2 and its recruitment to stress granules and

P-bodies. Mol Biol Cell 20: 3273-3284.

153. Barber GN (2009) The NFAR's (nuclear factors associated with dsRNA):

evolutionarily conserved members of the dsRNA binding protein family. RNA

Biol 6: 35-39.

154. Laraki G, Clerzius G, Daher A, Melendez-Pena C, Daniels S, et al. (2008)

Interactions between the double-stranded RNA-binding proteins TRBP and

173

PACT define the Medipal domain that mediates protein-protein interactions.

RNA Biol 5: 92-103.

155. Murphy D, Dancis B, Brown JR (2008) The evolution of core proteins involved

in microRNA biogenesis. BMC Evol Biol 8: 92.

156. Kok KH, Ng MH, Ching YP, Jin DY (2007) Human TRBP and PACT directly

interact with each other and associate with dicer to facilitate the production of

small interfering RNA. J Biol Chem 282: 17649-17657.

157. Lee Y, Hur I, Park SY, Kim YK, Suh MR, et al. (2006) The role of PACT in the

RNA silencing pathway. Embo J 25: 522-532.

158. Myers JW, Chi JT, Gong D, Schaner ME, Brown PO, et al. (2006) Minimizing

off-target effects by using diced siRNAs for RNA interference. J RNAi Gene

Silencing 2: 181-194.

159. Yuan YR, Pei Y, Ma JB, Kuryavyi V, Zhadina M, et al. (2005) Crystal structure

of A. aeolicus argonaute, a site-specific DNA-guided endoribonuclease,

provides insights into RISC-mediated mRNA cleavage. Mol Cell 19: 405-419.

160. Yuan YR, Pei Y, Chen HY, Tuschl T, Patel DJ (2006) A potential protein-RNA

recognition event along the RISC-loading pathway from the structure of A.

aeolicus Argonaute with externally bound siRNA. Structure 14: 1557-1565.

161. Wang Y, Sheng G, Juranek S, Tuschl T, Patel DJ (2008) Structure of the guide-

strand-containing argonaute silencing complex. Nature 456: 209-213.

162. Parker JS, Roe SM, Barford D (2005) Structural insights into mRNA recognition

from a PIWI domain-siRNA guide complex. Nature 434: 663-666.

163. Parker JS, Roe SM, Barford D (2004) Crystal structure of a PIWI protein

suggests mechanisms for siRNA recognition and slicer activity. Embo J 23:

4727-4737.

164. Okamura K, Ishizuka A, Siomi H, Siomi MC (2004) Distinct roles for Argonaute

proteins in small RNA-directed RNA cleavage pathways. Genes Dev 18: 1655-

1666.

165. Rand TA, Petersen S, Du F, Wang X (2005) Argonaute2 cleaves the anti-guide

strand of siRNA during RISC activation. Cell 123: 621-629.

174

166. Peters L, Meister G (2007) Argonaute proteins: mediators of RNA silencing. Mol

Cell 26: 611-623.

167. Hammond SM, Bernstein E, Beach D, Hannon GJ (2000) An RNA-directed

nuclease mediates post-transcriptional gene silencing in Drosophila cells.

Nature 404: 293-296.

168. Liu J, Carmell MA, Rivas FV, Marsden CG, Thomson JM, et al. (2004)

Argonaute2 is the catalytic engine of mammalian RNAi. Science 305: 1437-

1441.

169. Tomari Y, Matranga C, Haley B, Martinez N, Zamore PD (2004) A protein

sensor for siRNA asymmetry. Science 306: 1377-1380.

170. Hutvagner G (2005) Small RNA asymmetry in RNAi: function in RISC assembly

and gene regulation. FEBS Lett 579: 5850-5857.

171. Lin SL, Chang D, Ying SY (2005) Asymmetry of intronic pre-miRNA structures

in functional RISC assembly. Gene 356: 32-38.

172. Khvorova A, Reynolds A, Jayasena SD (2003) Functional siRNAs and miRNAs

exhibit strand bias. Cell 115: 209-216.

173. Kim K, Lee YS, Carthew RW (2007) Conversion of pre-RISC to holo-RISC by

Ago2 during assembly of RNAi complexes. Rna 13: 22-29.

174. Leuschner PJ, Ameres SL, Kueng S, Martinez J (2006) Cleavage of the siRNA

passenger strand during RISC assembly in human cells. EMBO Rep 7: 314-

320.

175. Miyoshi K, Tsukumo H, Nagami T, Siomi H, Siomi MC (2005) Slicer function

of Drosophila Argonautes and its involvement in RISC formation. Genes Dev

19: 2837-2848.

176. Matranga C, Tomari Y, Shin C, Bartel DP, Zamore PD (2005) Passenger-strand

cleavage facilitates assembly of siRNA into Ago2-containing RNAi enzyme

complexes. Cell 123: 607-620.

177. Tomari Y, Du T, Zamore PD (2007) Sorting of Drosophila small silencing

RNAs. Cell 130: 299-308.

178. Leuschner PJ, Obernosterer G, Martinez J (2005) MicroRNAs: Loquacious

speaks out. Curr Biol 15: R603-605.

175

179. Forstemann K, Tomari Y, Du T, Vagin VV, Denli AM, et al. (2005) Normal

microRNA maturation and germ-line stem cell maintenance requires

Loquacious, a double-stranded RNA-binding domain protein. PLoS Biol 3:

e236.

180. Saito K, Ishizuka A, Siomi H, Siomi MC (2005) Processing of pre-microRNAs

by the Dicer-1-Loquacious complex in Drosophila cells. PLoS Biol 3: e235.

181. Kanellopoulou C, Muljo SA, Kung AL, Ganesan S, Drapkin R, et al. (2005)

Dicer-deficient mouse embryonic stem cells are defective in differentiation and

centromeric silencing. Genes Dev 19: 489-501.

182. Forstemann K, Horwich MD, Wee L, Tomari Y, Zamore PD (2007) Drosophila

microRNAs are sorted into functionally distinct argonaute complexes after

production by dicer-1. Cell 130: 287-297.

183. Lazzaretti D, Tournier I, Izaurralde E (2009) The C-terminal domains of human

TNRC6A, TNRC6B, and TNRC6C silence bound transcripts independently of

Argonaute proteins. Rna 15: 1059-1066.

184. Fabian MR, Mathonnet G, Sundermeier T, Mathys H, Zipprich JT, et al. (2009)

Mammalian miRNA RISC recruits CAF1 and PABP to affect PABP-

dependent deadenylation. Mol Cell 35: 868-880.

185. Eulalio A, Tritschler F, Izaurralde E (2009) The GW182 protein family in animal

cells: new insights into domains required for miRNA-mediated gene silencing.

Rna 15: 1433-1442.

186. Eulalio A, Tritschler F, Buttner R, Weichenrieder O, Izaurralde E, et al. (2009)

The RRM domain in GW182 proteins contributes to miRNA-mediated gene

silencing. Nucleic Acids Res 37: 2974-2983.

187. Eulalio A, Helms S, Fritzsch C, Fauser M, Izaurralde E (2009) A C-terminal

silencing domain in GW182 is essential for miRNA function. Rna 15: 1067-

1077.

188. Ding XC, Grosshans H (2009) Repression of C. elegans microRNA targets at the

initiation level of translation requires GW182 proteins. Embo J 28: 213-222.

189. Chen CY, Zheng D, Xia Z, Shyu AB (2009) Ago-TNRC6 triggers microRNA-

mediated decay by promoting two deadenylation steps. Nat Struct Mol Biol.

176

190. Chekulaeva M, Filipowicz W, Parker R (2009) Multiple independent domains of

dGW182 function in miRNA-mediated repression in Drosophila. Rna 15: 794-

803.

191. Baillat D, Shiekhattar R (2009) Functional dissection of the human TNRC6

(GW182-related) family of proteins. Mol Cell Biol 29: 4144-4155.

192. Li S, Lian SL, Moser JJ, Fritzler ML, Fritzler MJ, et al. (2008) Identification of

GW182 and its novel isoform TNGW1 as translational repressors in Ago2-

mediated silencing. J Cell Sci 121: 4134-4144.

193. Eulalio A, Huntzinger E, Izaurralde E (2008) GW182 interaction with Argonaute

is essential for miRNA-mediated translational repression and mRNA decay.

Nat Struct Mol Biol 15: 346-353.

194. Liu J, Rivas FV, Wohlschlegel J, Yates JR, 3rd, Parker R, et al. (2005) A role for

the P-body component GW182 in microRNA function. Nat Cell Biol 7: 1261-

1266.

195. Ding L, Spencer A, Morita K, Han M (2005) The developmental timing regulator

AIN-1 interacts with miRISCs and may target the argonaute protein ALG-1 to

cytoplasmic P bodies in C. elegans. Mol Cell 19: 437-447.

196. Brennecke J, Stark A, Russell RB, Cohen SM (2005) Principles of microRNA-

target recognition. PLoS Biol 3: e85.

197. Enright AJ, John B, Gaul U, Tuschl T, Sander C, et al. (2003) MicroRNA targets

in Drosophila. Genome Biol 5: R1.

198. John B, Enright AJ, Aravin A, Tuschl T, Sander C, et al. (2004) Human

MicroRNA targets. PLoS Biol 2: e363.

199. Krek A, Grun D, Poy MN, Wolf R, Rosenberg L, et al. (2005) Combinatorial

microRNA target predictions. Nat Genet 37: 495-500.

200. Easow G, Teleman AA, Cohen SM (2007) Isolation of microRNA targets by

miRNP immunopurification. Rna 13: 1198-1204.

201. Lytle JR, Yario TA, Steitz JA (2007) Target mRNAs are repressed as efficiently

by microRNA-binding sites in the 5' UTR as in the 3' UTR. Proc Natl Acad Sci

U S A 104: 9667-9672.

177

202. Stark A, Lin MF, Kheradpour P, Pedersen JS, Parts L, et al. (2007) Discovery of

functional elements in 12 Drosophila genomes using evolutionary signatures.

Nature 450: 219-232.

203. Gu S, Jin L, Zhang F, Sarnow P, Kay MA (2009) Biological basis for restriction

of microRNA targets to the 3' untranslated region in mammalian mRNAs. Nat

Struct Mol Biol 16: 144-150.

204. Xie X, Lu J, Kulbokas EJ, Golub TR, Mootha V, et al. (2005) Systematic

discovery of regulatory motifs in human promoters and 3' UTRs by

comparison of several mammals. Nature 434: 338-345.

205. Vella MC, Choi EY, Lin SY, Reinert K, Slack FJ (2004) The C. elegans

microRNA let-7 binds to imperfect let-7 complementary sites from the lin-41

3'UTR. Genes Dev 18: 132-137.

206. Lim LP, Lau NC, Weinstein EG, Abdelhakim A, Yekta S, et al. (2003) The

microRNAs of Caenorhabditis elegans. Genes Dev 17: 991-1008.

207. Wang Y, Juranek S, Li H, Sheng G, Wardle GS, et al. (2009) Nucleation,

propagation and cleavage of target RNAs in Ago silencing complexes. Nature

461: 754-761.

208. Sethupathy P, Megraw M, Hatzigeorgiou AG (2006) A guide through present

computational approaches for the identification of mammalian microRNA

targets. Nat Methods 3: 881-886.

209. Kertesz M, Iovino N, Unnerstall U, Gaul U, Segal E (2007) The role of site

accessibility in microRNA target recognition. Nat Genet 39: 1278-1284.

210. Miranda KC, Huynh T, Tay Y, Ang YS, Tam WL, et al. (2006) A pattern-based

method for the identification of MicroRNA binding sites and their

corresponding heteroduplexes. Cell 126: 1203-1217.

211. Hall TM (2005) Structure and function of argonaute proteins. Structure 13: 1403-

1408.

212. Song JJ, Smith SK, Hannon GJ, Joshua-Tor L (2004) Crystal structure of

Argonaute and its implications for RISC slicer activity. Science 305: 1434-

1437.

178

213. Meister G, Tuschl T (2004) Mechanisms of gene silencing by double-stranded

RNA. Nature 431: 343-349.

214. Bhattacharyya SN, Habermacher R, Martine U, Closs EI, Filipowicz W (2006)

Relief of microRNA-mediated translational repression in human cells

subjected to stress. Cell 125: 1111-1124.

215. Chu CY, Rana TM (2006) Translation repression in human cells by microRNA-

induced gene silencing requires RCK/p54. PLoS Biol 4: e210.

216. Liu J, Valencia-Sanchez MA, Hannon GJ, Parker R (2005) MicroRNA-

dependent localization of targeted mRNAs to mammalian P-bodies. Nat Cell

Biol 7: 719-723.

217. Zeng Y, Cullen BR (2003) Sequence requirements for micro RNA processing

and function in human cells. Rna 9: 112-123.

218. Zeng Y, Wagner EJ, Cullen BR (2002) Both natural and designed micro RNAs

can inhibit the expression of cognate mRNAs when expressed in human cells.

Mol Cell 9: 1327-1333.

219. Decker CJ, Parker R (1994) Mechanisms of mRNA degradation in eukaryotes.

Trends Biochem Sci 19: 336-340.

220. Wormington M (1993) Poly(A) and translation: development control. Curr Opin

Cell Biol 5: 950-954.

221. Kiriakidou M, Tan GS, Lamprinaki S, De Planell-Saguer M, Nelson PT, et al.

(2007) An mRNA m7G cap binding-like motif within human Ago2 represses

translation. Cell 129: 1141-1151.

222. Mathonnet G, Fabian MR, Svitkin YV, Parsyan A, Huck L, et al. (2007)

MicroRNA inhibition of translation initiation in vitro by targeting the cap-

binding complex eIF4F. Science 317: 1764-1767.

223. Thermann R, Hentze MW (2007) Drosophila miR2 induces pseudo-polysomes

and inhibits translation initiation. Nature 447: 875-878.

224. Chendrimada TP, Finn KJ, Ji X, Baillat D, Gregory RI, et al. (2007) MicroRNA

silencing through RISC recruitment of eIF6. Nature 447: 823-828.

225. Wang B, Yanez A, Novina CD (2008) MicroRNA-repressed mRNAs contain 40S

but not 60S components. Proc Natl Acad Sci U S A 105: 5343-5348.

179

226. Sanvito F, Piatti S, Villa A, Bossi M, Lucchini G, et al. (1999) The beta4 integrin

interactor p27(BBP/eIF6) is an essential nuclear matrix protein involved in 60S

ribosomal subunit assembly. J Cell Biol 144: 823-837.

227. Ceci M, Gaviraghi C, Gorrini C, Sala LA, Offenhauser N, et al. (2003) Release of

eIF6 (p27BBP) from the 60S subunit allows 80S ribosome assembly. Nature

426: 579-584.

228. Nottrott S, Simard MJ, Richter JD (2006) Human let-7a miRNA blocks protein

production on actively translating polyribosomes. Nat Struct Mol Biol 13:

1108-1114.

229. Petersen CP, Bordeleau ME, Pelletier J, Sharp PA (2006) Short RNAs repress

translation after initiation in mammalian cells. Mol Cell 21: 533-542.

230. Maroney PA, Yu Y, Fisher J, Nilsen TW (2006) Evidence that microRNAs are

associated with translating messenger RNAs in human cells. Nat Struct Mol

Biol 13: 1102-1107.

231. Jacobson A, Peltz SW (1996) Interrelationships of the pathways of mRNA decay

and translation in eukaryotic cells. Annu Rev Biochem 65: 693-739.

232. Kruys V, Marinx O, Shaw G, Deschamps J, Huez G (1989) Translational

blockade imposed by cytokine-derived UA-rich sequences. Science 245: 852-

855.

233. Kruys V, Wathelet M, Poupart P, Contreras R, Fiers W, et al. (1987) The 3'

untranslated region of the human interferon-beta mRNA has an inhibitory

effect on translation. Proc Natl Acad Sci U S A 84: 6030-6034.

234. Kruys VI, Wathelet MG, Huez GA (1988) Identification of a translation

inhibitory element (TIE) in the 3' untranslated region of the human interferon-

beta mRNA. Gene 72: 191-200.

235. Wu L, Fan J, Belasco JG (2008) Importance of translation and nonnucleolytic

ago proteins for on-target RNA interference. Curr Biol 18: 1327-1332.

236. Aleman LM, Doench J, Sharp PA (2007) Comparison of siRNA-induced off-

target RNA and protein effects. Rna 13: 385-395.

180

237. Standart N, Jackson RJ (2007) MicroRNAs repress translation of m7Gppp-

capped target mRNAs in vitro by inhibiting initiation and promoting

deadenylation. Genes Dev 21: 1975-1982.

238. Kong YW, Cannell IG, de Moor CH, Hill K, Garside PG, et al. (2008) The

mechanism of micro-RNA-mediated translation repression is determined by

the promoter of the target gene. Proc Natl Acad Sci U S A 105: 8866-8871.

239. Hendrickson DG, Hogan DJ, Herschlag D, Ferrell JE, Brown PO (2008)

Systematic identification of mRNAs recruited to argonaute 2 by specific

microRNAs and corresponding changes in transcript abundance. PLoS ONE 3:

e2126.

240. Grishok A, Pasquinelli AE, Conte D, Li N, Parrish S, et al. (2001) Genes and

mechanisms related to RNA interference regulate expression of the small

temporal RNAs that control C. elegans developmental timing. Cell 106: 23-34.

241. Haley B, Zamore PD (2004) Kinetic analysis of the RNAi enzyme complex. Nat

Struct Mol Biol 11: 599-606.

242. Lai EC (2002) Micro RNAs are complementary to 3' UTR sequence motifs that

mediate negative post-transcriptional regulation. Nat Genet 30: 363-364.

243. Wang Y, Juranek S, Li H, Sheng G, Tuschl T, et al. (2008) Structure of an

argonaute silencing complex with a seed-containing guide DNA and target

RNA duplex. Nature 456: 921-926.

244. Rashid UJ, Paterok D, Koglin A, Gohlke H, Piehler J, et al. (2007) Structure of

Aquifex aeolicus argonaute highlights conformational flexibility of the PAZ

domain as a potential regulator of RNA-induced silencing complex function. J

Biol Chem 282: 13824-13832.

245. Blaszczyk J, Tropea JE, Bubunenko M, Routzahn KM, Waugh DS, et al. (2001)

Crystallographic and modeling studies of RNase III suggest a mechanism for

double-stranded RNA cleavage. Structure (Camb) 9: 1225-1236.

246. Tomari Y, Du T, Haley B, Schwarz DS, Bennett R, et al. (2004) RISC assembly

defects in the Drosophila RNAi mutant armitage. Cell 116: 831-841.

181

247. Cook HA, Koppetsch BS, Wu J, Theurkauf WE (2004) The Drosophila SDE3

homolog armitage is required for oskar mRNA silencing and embryonic axis

specification. Cell 116: 817-829.

248. Pellino JL, Jaskiewicz L, Filipowicz W, Sontheimer EJ (2005) ATP modulates

siRNA interactions with an endogenous human Dicer complex. Rna 11: 1719-

1724.

249. Pellino JL, Sontheimer EJ (2003) R2D2 leads the silencing trigger to mRNA's

death star. Cell 115: 132-133.

250. Doi N, Zenno S, Ueda R, Ohki-Hamazaki H, Ui-Tei K, et al. (2003) Short-

interfering-RNA-mediated gene silencing in mammalian cells requires Dicer

and eIF2C translation initiation factors. Curr Biol 13: 41-46.

251. Caudy AA, Myers M, Hannon GJ, Hammond SM (2002) Fragile X-related

protein and VIG associate with the RNA interference machinery. Genes Dev

16: 2491-2496.

252. Carthew RW (2006) Gene regulation by microRNAs. Curr Opin Genet Dev 16:

203-208.

253. Chen PY, Meister G (2005) microRNA-guided posttranscriptional gene

regulation. Biol Chem 386: 1205-1218.

254. Didiano D, Hobert O (2006) Perfect seed pairing is not a generally reliable

predictor for miRNA-target interactions. Nat Struct Mol Biol 13: 849-851.

255. SFGF (2008) Human Exonic Evidence Based Oligonucleotide (HEEBO) Arrays

256. Ambion (2008) mirVana™ miRNA Probe Set

257. Ferea TL, Brown PO (1999) Observing the living genome. Curr Opin Genet Dev

9: 715-722.

258. Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays

applied to the ionizing radiation response. Proc Natl Acad Sci U S A 98: 5116-

5121.

259. Efron B, Tibshirani R (2002) Empirical bayes methods and false discovery rates

for microarrays. Genet Epidemiol 23: 70-86.

182

260. Bailey TL, Elkan C (1994) Fitting a mixture model by expectation maximization

to discover motifs in biopolymers. Proc Int Conf Intell Syst Mol Biol 2: 28-36.

261. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, et al. (2005)

Gene set enrichment analysis: a knowledge-based approach for interpreting

genome-wide expression profiles. Proc Natl Acad Sci U S A 102: 15545-

15550.

262. Beitzinger M, Peters L, Zhu JY, Kremmer E, Meister G (2007) Identification of

Human microRNA Targets From Isolated Argonaute Protein Complexes. RNA

Biol 4.

263. Karginov FV, Conaco C, Xuan Z, Schmidt BH, Parker JS, et al. (2007) A

biochemical approach to identifying microRNA targets. Proc Natl Acad Sci U

S A 104: 19291-19296.

264. Zhang L, Ding L, Cheung TH, Dong MQ, Chen J, et al. (2007) Systematic

identification of C. elegans miRISC proteins, miRNAs, and mRNA targets by

their interactions with GW182 proteins AIN-1 and AIN-2. Mol Cell 28: 598-

613.

265. Fare TL, Coffey EM, Dai H, He YD, Kessler DA, et al. (2003) Effects of

atmospheric ozone on microarray data quality. Anal Chem 75: 4672-4675.

266. Demeter J, Beauheim C, Gollub J, Hernandez-Boussard T, Jin H, et al. (2007)

The Stanford Microarray Database: implementation of new analysis tools and

open source release of software. Nucleic Acids Res 35: D766-770.

267. Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and

display of genome-wide expression patterns. Proc Natl Acad Sci U S A 95:

14863-14868.

268. Saldanha AJ (2004) Java Treeview--extensible visualization of microarray data.

Bioinformatics 20: 3246-3248.

269. Hosack DA, Dennis G, Jr., Sherman BT, Lane HC, Lempicki RA (2003)

Identifying biological themes within lists of genes with EASE. Genome Biol 4:

R70.

183

270. Seggerson K, Tang L, Moss EG (2002) Two genetic circuits repress the

Caenorhabditis elegans heterochronic gene lin-28 after translation initiation.

Dev Biol 243: 215-225.

271. Slack F, Ruvkun G (1997) Temporal pattern formation by heterochronic genes.

Annu Rev Genet 31: 611-634.

272. Bachvarova RF (1992) A maternal tail of poly(A): the long and the short of it.

Cell 69: 895-897.

273. Deshpande AK, Chatterjee B, Roy AK (1979) Translation and stability of rat

liver messenger RNA for alpha 2 mu-globulin in Xenopus oocyte. The role of

terminal poly(A). J Biol Chem 254: 8937-8942.

274. Drummond DR, Armstrong J, Colman A (1985) The effect of capping and

polyadenylation on the stability, movement and translation of synthetic

messenger RNAs in Xenopus oocytes. Nucleic Acids Res 13: 7375-7394.

275. Gallie DR (1991) The cap and poly(A) tail function synergistically to regulate

mRNA translational efficiency. Genes Dev 5: 2108-2116.

276. Gallie DR, Lucas WJ, Walbot V (1989) Visualizing mRNA expression in plant

protoplasts: factors influencing efficient mRNA uptake and translation. Plant

Cell 1: 301-311.

277. Wang B, Love TM, Call ME, Doench JG, Novina CD (2006) Recapitulation of

short RNA-directed translational gene silencing in vitro. Mol Cell 22: 553-560.

278. Ikeda K, Satoh M, Pauley KM, Fritzler MJ, Reeves WH, et al. (2006) Detection

of the argonaute protein Ago2 and microRNAs in the RNA induced silencing

complex (RISC) using a monoclonal antibody. J Immunol Methods 317: 38-

44.

279. Forman JJ, Legesse-Miller A, Coller HA (2008) A search for conserved

sequences in coding regions reveals that the let-7 microRNA targets Dicer

within its coding sequence. Proc Natl Acad Sci U S A 105: 14879-14884.

280. Tay Y, Zhang J, Thomson AM, Lim B, Rigoutsos I (2008) MicroRNAs to

Nanog, Oct4 and Sox2 coding regions modulate embryonic stem cell

differentiation. Nature 455: 1124-1128.

184

281. Arava Y, Wang Y, Storey JD, Liu CL, Brown PO, et al. (2003) Genome-wide

analysis of mRNA translation profiles in Saccharomyces cerevisiae. Proc Natl

Acad Sci U S A 100: 3889-3894.

282. Wolin SL, Walter P (1988) Ribosome pausing and stacking during translation of

a eukaryotic mRNA. Embo J 7: 3559-3569.

283. Arava Y, Boas FE, Brown PO, Herschlag D (2005) Dissecting eukaryotic

translation and its control by ribosome density mapping. Nucleic Acids Res 33:

2421-2432.

284. Ingolia NT, Ghaemmaghami S, Newman JR, Weissman JS (2009) Genome-wide

analysis in vivo of translation with nucleotide resolution using ribosome

profiling. Science 324: 218-223.

285. Richter JD (1999) Cytoplasmic polyadenylation in development and beyond.

Microbiol Mol Biol Rev 63: 446-456.

286. Wormington M (1994) Unmasking the role of the 3' UTR in the cytoplasmic

polyadenylation and translational regulation of maternal mRNAs. Bioessays

16: 533-535.

287. Hegde RS, Kang SW (2008) The concept of translocational regulation. J Cell

Biol 182: 225-232.

288. Fang S, Jensen JP, Ludwig RL, Vousden KH, Weissman AM (2000) Mdm2 is a

RING finger-dependent ubiquitin protein ligase for itself and p53. J Biol Chem

275: 8945-8951.

289. Ito K, Adachi S, Iwakami R, Yasuda H, Muto Y, et al. (2001) N-Terminally

extended human ubiquitin-conjugating enzymes (E2s) mediate the

ubiquitination of RING-finger proteins, ARA54 and RNF8. Eur J Biochem

268: 2725-2732.

290. Yang Y, Fang S, Jensen JP, Weissman AM, Ashwell JD (2000) Ubiquitin protein

ligase activity of IAPs and their degradation in proteasomes in response to

apoptotic stimuli. Science 288: 874-877.

291. Marson A, Levine SS, Cole MF, Frampton GM, Brambrink T, et al. (2008)

Connecting microRNA genes to the core transcriptional regulatory circuitry of

embryonic stem cells. Cell 134: 521-533.

185

292. Landthaler M, Yalcin A, Tuschl T (2004) The human DiGeorge syndrome critical

region gene 8 and Its D. melanogaster homolog are required for miRNA

biogenesis. Curr Biol 14: 2162-2167.

293. Holtz J, Pasquinelli AE (2009) Uncoupling of lin-14 mRNA and protein

repression by nutrient deprivation in Caenorhabditis elegans. Rna 15: 400-405.

294. Coller J, Parker R (2005) General translational repression by activators of mRNA

decapping. Cell 122: 875-886.

295. Kawai T, Fan J, Mazan-Mamczarz K, Gorospe M (2004) Global mRNA

stabilization preferentially linked to translational repression during the

endoplasmic reticulum stress response. Mol Cell Biol 24: 6773-6787.

296. Sheth U, Parker R (2003) Decapping and decay of messenger RNA occur in

cytoplasmic processing bodies. Science 300: 805-808.

297. Schwartz DC, Parker R (1999) Mutations in translation initiation factors lead to

increased rates of deadenylation and decapping of mRNAs in Saccharomyces

cerevisiae. Mol Cell Biol 19: 5247-5256.

298. Bandyopadhyay R, Coutts M, Krowczynska A, Brawerman G (1990) Nuclease

activity associated with mammalian mRNA in its native state: possible basis

for selectivity in mRNA decay. Mol Cell Biol 10: 2060-2069.

299. Caruccio N, Ross J (1994) Purification of a human polyribosome-associated 3' to

5' exoribonuclease. J Biol Chem 269: 31814-31821.

300. Grafi G, Sela I, Galili G (1993) Translational regulation of human beta interferon

mRNA: association of the 3' AU-rich sequence with the poly(A) tail reduces

translation efficiency in vitro. Mol Cell Biol 13: 3487-3493.

301. Marinx O, Bertrand S, Karsenti E, Huez G, Kruys V (1994) Fertilization of

Xenopus eggs imposes a complete translational arrest of mRNAs containing

3'UUAUUUAU elements. FEBS Lett 345: 107-112.

302. Muhlrad D, Decker CJ, Parker R (1995) Turnover mechanisms of the stable yeast

PGK1 mRNA. Mol Cell Biol 15: 2145-2156.

303. Klapholz-Brown Z, Walmsley GG, Nusse YM, Nusse R, Brown PO (2007)

Transcriptional program induced by Wnt protein in human fibroblasts suggests

186

mechanisms for cell cooperativity in defining tissue microenvironments. PLoS

ONE 2: e945.

304. Backes C, Keller A, Kuentzer J, Kneissl B, Comtesse N, et al. (2007) GeneTrail--

advanced gene set enrichment analysis. Nucleic Acids Res 35: W186-192.

305. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J

Statist 6: 65–70

306. Mayr C, Bartel DP (2009) Widespread shortening of 3'UTRs by alternative

cleavage and polyadenylation activates oncogenes in cancer cells. Cell 138:

673-684.

307. Sandberg R, Neilson JR, Sarma A, Sharp PA, Burge CB (2008) Proliferating

cells express mRNAs with shortened 3' untranslated regions and fewer

microRNA target sites. Science 320: 1643-1647.

308. Licatalosi DD, Mele A, Fak JJ, Ule J, Kayikci M, et al. (2008) HITS-CLIP yields

genome-wide insights into brain alternative RNA processing. Nature 456: 464-

469.

309. Chi SW, Zang JB, Mele A, Darnell RB (2009) Argonaute HITS-CLIP decodes

microRNA-mRNA interaction maps. Nature 460: 479-486.

310. Goff LA, Davila J, Swerdel MR, Moore JC, Cohen RI, et al. (2009) Ago2

immunoprecipitation identifies predicted microRNAs in human embryonic

stem cells and neural precursors. PLoS One 4: e7192.

311. Tan LP, Seinen E, Duns G, de Jong D, Sibon OC, et al. (2009) A high throughput

experimental approach to identify miRNA targets in human cells. Nucleic

Acids Res.

312. Huang X, Ding L, Bennewith KL, Tong RT, Welford SM, et al. (2009) Hypoxia-

inducible mir-210 regulates normoxic gene expression involved in tumor

initiation. Mol Cell 35: 856-867.

313. Kadener S, Menet JS, Sugino K, Horwich MD, Weissbein U, et al. (2009) A role

for microRNAs in the Drosophila circadian clock. Genes Dev 23: 2179-2191.

187

314. Zhang L, Hammell M, Kudlow BA, Ambros V, Han M (2009) Systematic

analysis of dynamic miRNA-target interactions during C. elegans

development. Development 136: 3043-3055.

315. Yoo AS, Staahl BT, Chen L, Crabtree GR (2009) MicroRNA-mediated switching

of chromatin-remodelling complexes in neural development. Nature 460: 642-

646.

316. Zhou JY, Ma WL, Liang S, Zeng Y, Shi R, et al. (2009) Analysis of microRNA

expression profiles during the cell cycle in synchronized HeLa cells. BMB Rep

42: 593-598.

317. Hwang HW, Wentzel EA, Mendell JT (2007) A hexanucleotide element directs

microRNA nuclear import. Science 315: 97-100.

318. Landthaler M, Gaidatzis D, Rothballer A, Chen PY, Soll SJ, et al. (2008)

Molecular characterization of human Argonaute-containing ribonucleoprotein

complexes and their bound target mRNAs. Rna.

319. Meister G, Landthaler M, Peters L, Chen PY, Urlaub H, et al. (2005)

Identification of novel argonaute-associated proteins. Curr Biol 15: 2149-2155.

320. Beilharz TH, Humphreys DT, Clancy JL, Thermann R, Martin DI, et al. (2009)

microRNA-mediated messenger RNA deadenylation contributes to

translational repression in mammalian cells. PLoS One 4: e6783.

321. Elemento O, Slonim N, Tavazoie S (2007) A universal framework for regulatory

element discovery across all genomes and data types. Mol Cell 28: 337-350.

322. Cleveland WS (1979) Robust Locally Weighted Fitting and Smoothing

Scatterplots. Journal of the American Statistical Association 74: 829-836.